Homebrew PCBs

Is there a website that describes rules for layout that will ensure 
adequate spacing for pick and place machines?

I'm doing a few by hand but if the project is successful I might 
have a shot a selling a thousand or so. Can't see myself sitting in 
the basement soldering finelines that long...

tnx
jtm

jtm,

Don't have a website for you, but do have a couple suggestions. 
Instead of looking for a "general" solution: Talk to the folks you 
would have (potentially) doing this for you...

There are a wide variety of pick and places being used, from older 
models to the latest/greatest; and the "rules" will be different for 
each.

In addition, the number of trays and reels that the stuffing house 
machine(s) can handle may be a further influence on your design...

Sometimes using two or even three parts of a type that you "already" 
are using elsewhere on the board, ends up being less expensive than 
using the "one" part which "can do" what the three parts replicate. 
This is because you are ordering larger numbers of fewer parts, and 
the pick and place speed in placing THREE parts will make up for the 
need to place ONE "weird" part... (That may require a second run, 
because the reels/bins are all used up..)

But the folks you want answering this are the ones you're going to 
trust to populate and solder the boards for you, IMO...

Hope this helps,

Ballendo 

P.S. You'll also need them to tell you where they expect different 
parts to be "picked up"... Not all machines work the same way... 



--- In Homebrew_PCBs@yahoogroups.com, "mdgolfbum" <jim@j...> wrote:

> Is there a website that describes rules for layout that will ensure 
> adequate spacing for pick and place machines?
> 
> I'm doing a few by hand but if the project is successful I might 
> have a shot a selling a thousand or so. Can't see myself sitting in 
> the basement soldering finelines that long...
> 
> tnx
> jtm

--- In Homebrew_PCBs@yahoogroups.com, "mdgolfbum" <jim@j...> wrote:
> Is there a website that describes rules for layout that will ensure 
> adequate spacing for pick and place machines?
> 
> I'm doing a few by hand but if the project is successful I might 
> have a shot a selling a thousand or so. Can't see myself sitting in 
> the basement soldering finelines that long...
> 
> tnx
> jtm


I am in the same boat.

I have about 10 boards with smt parts that I would like to get done.  
The next bunch will hopefully be 100 boards and then after that, 
1,000 boards.

I did a little checking and found one guy who will do boards for 
something about 25 cents per SMT part and 75 cents per thru-hole 
part.  yup, a resistor is a thru-hole and yes, I pay 2 cents for them 
too.


did more checking...

Outsource to China and the price structure changed to $200 set-up, 
$250 engineering charge and two cents per part.  that is much better 
for the 1,000 board project than the 10 board or 100 board ones.

I also just started looking into silkscreen or template for screening 
on solder paste for toaster oven soldering.  That starts at about 
$75.00 for a metal, laser cut stencil, or I can cut the solder pads 
our of plastic with an exacto knife.  

Anyone know the thickness for solder paste ?

Alternately, a solder paste dispenser would be a great addition to 
the shop.

Dave

(Ballendo, is your e-mail working yet ?  I've sent you some responses 
to your questions, but have not heard back)

On Wed, 21 Apr 2004, Dave Mucha wrote:

>
> I have about 10 boards with smt parts that I would like to get done.
> The next bunch will hopefully be 100 boards and then after that,
> 1,000 boards.
>

Hi Dave,

  I don't know if this may help you:

  http://www.pad2pad.com/company/index.php
  http://www.olimex.com


best regards,
Vasile
http://surducan.netfirms.com

----- Original Message ----- 

From: "Dave Mucha" <dave_mucha@...>
To: <Homebrew_PCBs@yahoogroups.com>
Sent: Wednesday, April 21, 2004 1:24 PM
Subject: [Homebrew_PCBs] Re: Pick and place rules for layout?


> --- In Homebrew_PCBs@yahoogroups.com, "mdgolfbum" <jim@j...> wrote:
> > Is there a website that describes rules for layout that will ensure 
> > adequate spacing for pick and place machines?
> > 
> > I'm doing a few by hand but if the project is successful I might 
> > have a shot a selling a thousand or so. Can't see myself sitting in 
> > the basement soldering finelines that long...
> > 
> > tnx
> > jtm
> 
> 
> I am in the same boat.
> 
> I have about 10 boards with smt parts that I would like to get done.  
> The next bunch will hopefully be 100 boards and then after that, 
> 1,000 boards.
> 
> I did a little checking and found one guy who will do boards for 
> something about 25 cents per SMT part and 75 cents per thru-hole 
> part.  yup, a resistor is a thru-hole and yes, I pay 2 cents for them 
> too.
> 
> 
> did more checking...
> 
> Outsource to China and the price structure changed to $200 set-up, 
> $250 engineering charge and two cents per part.  that is much better 
> for the 1,000 board project than the 10 board or 100 board ones.
> 
> I also just started looking into silkscreen or template for screening 
> on solder paste for toaster oven soldering.  That starts at about 
> $75.00 for a metal, laser cut stencil, or I can cut the solder pads 
> our of plastic with an exacto knife.  
> 
> Anyone know the thickness for solder paste ?
> 
> Alternately, a solder paste dispenser would be a great addition to 
> the shop.

This is a useful page:

http://www.pcbexpress.com/stencils/apply_solder_paste.php

The thickness is determined by the stencil thickness.

Leon

--- Leon Heller <leon_heller@...> wrote:

> > Alternately, a solder paste dispenser would be a
> great addition to 
> > the shop.

Why not etch your own stencil using fecl, cucl or
whatever etchant you already have? You could take a
piece of thin copper, photo sensitive it with an image
of your solder pads, develop and then etch. It's gotta
be a lot cheaper than that kit.

Regards

Marvin

=====
Registered Linux User No. 80253
If you use linux, get counted at: 
http://www.linuxcounter.org


	
		
__________________________________
Do you Yahoo!?
Yahoo! Photos: High-quality 4x6 digital prints for 25\ufffd
http://photos.yahoo.com/ph/print_splash

> 
>   http://www.pad2pad.com/company/index.php

I looked at Pad2Pad, but they are similar to Express.  As long as 
your parts fit their structure, you are allowed to use them.

I have some 57 dia holes with 80 mil pads.  That is just not going to 
work with these cookie cutter houses.

Also, their 0805 SMT pads are huge and don't leave any room for 
traces between what the industry standard size would allow.

also, the software has a hard time with traces and nets and even 
traces too close per design rules.  

It is just too young, maybe in another few months it will be better.


>   http://www.olimex.com


I'm a few day away from getting approval on the Gerbers for getting 
quotes from people like Olimex.

I'm putting 4 seperate type of board on one panel and many of hte 
board houses I have contacted have been more than willing to do the 
step and repeat for me, others flatly refuse.

Olimex is one of the very few that have offered to cut the panels 
into smaller sections for no charge.

But, I need them in a large panel to do the soldering.

Seems there is always some little problem with something.

Dave

> 
> 
> best regards,
> Vasile
> http://surducan.netfirms.com

On Wed, 21 Apr 2004, Dave Mucha wrote:

>
> I'm putting 4 seperate type of board on one panel and many of hte
> board houses I have contacted have been more than willing to do the
> step and repeat for me, others flatly refuse.

  would you so kind and tell me which CAD are using ? I'm using Eagle
now, but it seems there is no panelisation possibilities... Copying and
paste all layers are forbidden.

>
> Olimex is one of the very few that have offered to cut the panels
> into smaller sections for no charge.
>
> But, I need them in a large panel to do the soldering.
>
> Seems there is always some little problem with something.

  I don't know how large "mass" production do you have, but till 1000
pieces I don't see any impediment in cutting boards "at home".

best,
Vasile
http://surducan.netfirms.com

use a gerber app.  I use Pentalogix Viewmate (free, google for 
site).  Use the CAM processor in Eagle to output gerber rs247x 
files.  You can import multiple gerbers into viewmate and combine 
them to make a panel.  

Phil

--- In Homebrew_PCBs@yahoogroups.com, Vasile Surducan <vasile@s...> 
wrote:
> 
> 
> 
> 
> On Wed, 21 Apr 2004, Dave Mucha wrote:
> 
> >
> > I'm putting 4 seperate type of board on one panel and many of hte
> > board houses I have contacted have been more than willing to do 
the
> > step and repeat for me, others flatly refuse.
> 
>   would you so kind and tell me which CAD are using ? I'm using 
Eagle
> now, but it seems there is no panelisation possibilities... Copying 
and
> paste all layers are forbidden.
> 
> >
> > Olimex is one of the very few that have offered to cut the panels
> > into smaller sections for no charge.
> >
> > But, I need them in a large panel to do the soldering.
> >
> > Seems there is always some little problem with something.
> 
>   I don't know how large "mass" production do you have, but till 
1000

> pieces I don't see any impediment in cutting boards "at home".
> 
> best,
> Vasile
> http://surducan.netfirms.com

> > $75.00 for a metal, laser cut stencil, or I can cut the solder pads
> > our of plastic with an exacto knife.

$75 is pretty cheap for stencils, well worth the laser cut ones, it will
save you much agravation. In the past, one could get chemcially etched
stencils, but they did not work anywhere near as well as the laser cut ones.

One key thing is you can put multiple pcb's on one stencil, back when we
were paying $600/stencil, we probably had 4-5 boards on each one to keep
costs low.

I doubt you could manually cut the stencils with an exacto knife to within a
few thousands. One usually plays some games with the stencil size to
optimize process. For a couple boards, probably the xacto might work, but
for even 10, you really need the precision, otherwise you pay for it in
rework time.

Used EFD machines are pretty cheap, and once you get the bugs out do very
well. Just write down the settings for each pad size and remember to reset
them for every footprint change.

Ron

Did some experimenting...

I have now etched away the copper of both welding rods, which nicely
saturated my etchant too...

1 rod is 8mm * 22cm i think which is 55 square cm.
i connected both rods together to get ~~100 square centimeters.


for the other electrode i used a copper wire, diameter about 4mm and 
length 10cm.
which is 12,5cm square.

i can only prevent chlorine gas by currents well below 0,5A (you know it 
is hard to tell when
there is really no gas).

- that would be 5mA / square cm at the carbon then.

However i think the area of the copper wire is too big, i can not get 
hydrogen at all, even
if there is severe chlorine. so i assumed it is too big and now i use a 
1.5mm^2 wire immersed
only 2 centimeters. now i get hydrogen before i get chlorine.
I have it running at 150mA right now, and will see if it works.

The copper plates like a sponge, and it is kind of hard to get it out 
without
shaking it all off again (and loosing it in the etchant).
It is funny how the copper grows. I suspect the area hugely increases when 
the first
copper is deposited and the current density is wrong again.

I must say right now it is not looking very well. the etchant is much 
darker than it should be and
all the experiments have not yet managed to regenerate it at all, I have 
only managed to produce
chlorine.



I have also another question about chlorine, what are those tablets made 
of for the swimming pool?
They say they are chlorine tablets but i am not sure. (thought pure 
chlorine only comes in gas form).
I wonder if one could use them to regenerate the bath, what i notice when 
i have too high a current
is that the chlorine gas produced clearly regenerates the etchant in the 
surroundings of the carbon electrodes
as it turns a lighter color.


I will try to optimize the process further but if it does not improove i 
will just stick with H2O2....

ST

Ron,

I interested in a couple things you bring up... Snips, Questions 
below>

>In Homebrew_PCBs, "Ron Amundson" <ron_amundson@h...> wrote:
> In the past, one could get chemically etched stencils, but they did 
> not work anywhere near as well as the laser cut ones.

What were the problems with the chemically etched compared to laser 
cut? IOW, please explain "work anywhere near as well"?

> For a couple boards, probably the xacto might work, but
> for even 10, you really need the precision, otherwise you pay for 
> it in rework time.

Hence the use of my 500-1000 buck CNC drill to CUT them cleanly, 
precisely, and accurately...

> Used EFD machines are pretty cheap<s>

Where?

Thank you in advance,

Ballendo

Stefan Trethan wrote:
> Did some experimenting...
> 
> I have now etched away the copper of both welding rods, which nicely
> saturated my etchant too...
> 
> 1 rod is 8mm * 22cm i think which is 55 square cm.
> i connected both rods together to get ~~100 square centimeters.
> 
> 
> for the other electrode i used a copper wire, diameter about 4mm and 
> length 10cm.
> which is 12,5cm square.
> 
> i can only prevent chlorine gas by currents well below 0,5A (you know it 
> is hard to tell when
> there is really no gas).
> 
> - that would be 5mA / square cm at the carbon then.

Excellent work, you now have a starting figure for the current density of 
graphite anodes. In practice you may need less current density to keep long 
life of the anode.

> However i think the area of the copper wire is too big, i can not get 
> hydrogen at all, even
> if there is severe chlorine. so i assumed it is too big and now i use a 
> 1.5mm^2 wire immersed
> only 2 centimeters. now i get hydrogen before i get chlorine.

Can I ask why are you trying to produce hydrogen gas at all ?
Are you saying that no visible signs of copper are plated without also 
producing hydrogen gas at the cathode ? If this is the case then you must 
have very little copper dissolved in the solution, such it would almost 
look very pale clear green.


> The copper plates like a sponge, and it is kind of hard to get it out 
> without
> shaking it all off again (and loosing it in the etchant).

that's correct. When metal and gas co-deposit you get metal sponge.

> 
> I have also another question about chlorine, what are those tablets made 
> of for the swimming pool?

calcium hypochlorite. This may regenerate CuCl2 etchant but I'm not sure 
the effect of byproduct, calcium chloride. It may not make a truly reusable 
etchant. commercially, they sometimes use sodium chlorate, which is similar 
except it is a more powerful oxidant, (more economical), and the byproduct 
is sodium chloride.

>
> Can I ask why are you trying to produce hydrogen gas at all ?
> Are you saying that no visible signs of copper are plated without also
> producing hydrogen gas at the cathode ? If this is the case then you must
> have very little copper dissolved in the solution, such it would almost
> look very pale clear green.
>

no, i get copper with no hydrogen too.
I aim for a area ratio that starts to produce hydrogen and chlorine at the 
same time,
and then use current slightly below that point.
It is better if i get hydrogen before chlorine because that means the 
copper area
is the absolute minimum. if i get clorine earlier then the copper could 
still be made
smaller.


> that's correct. When metal and gas co-deposit you get metal sponge.

correct, with lower current (no gas) i get no sponge, but a rough solid.
>
>>
>> I have also another question about chlorine, what are those tablets made
>> of for the swimming pool?
>
> calcium hypochlorite. This may regenerate CuCl2 etchant but I'm not sure
> the effect of byproduct, calcium chloride. It may not make a truly 
> reusable
> etchant. commercially, they sometimes use sodium chlorate, which is 
> similar
> except it is a more powerful oxidant, (more economical), and the 
> byproduct
> is sodium chloride.


i only ask because i have some leftover...
I might try it, but also might not. I guess it would not exactly decrease
clorine gas production.....


Yesterday i let it run for several hours with some 200mA and at then end i 
got visible regeneration.
I didn't want to run it overnight without supervision so i turned it off 
and pulled ou the copper.
now i will go and switch it on again i think.

The lid for the vertical tank seems to be set now, but i have still some 
work to do before i can
activate it.
It will hold 1 liter of etchant, and accept boards up to 35cm x 20cm which 
is more than i need.
It is also about the size i can make a TT (A4).
I'm stuck right now with creating the lid, it got more complicated since i 
aimed for electrolytic
regeneration. Right now i think i will seperate the lid in 3 sections, the 
middle conaining
the pcb clamp and the outer 2 containing the electrolysis electrodes. This 
way i can input/output
boards without disturbing electrolysis and i can also completely remove 
the electrodes if i should
once need the whole 35cm length. I will most likely make the copper 
electrode holding lid twice,
one holding a wire, the other blank, because obviously the copper needs to 
b taken out
if there is no current present.


I'll write again when there are new results or questions.


I would have one other question now - do you think the etchant penetrates 
the welding rods?
I connected them by drilling a hole and inserting a wire. then i sealed it 
with silicone.
If the etchant seeps through this is no good of course.

ST

On Thu, 22 Apr 2004 22:30:21 -0000, ballendo <ballendo@...> wrote:

> Ron,
>
> I interested in a couple things you bring up... Snips, Questions
> below>
>
>> In Homebrew_PCBs, "Ron Amundson" <ron_amundson@h...> wrote:
>> In the past, one could get chemically etched stencils, but they did
>> not work anywhere near as well as the laser cut ones.
>
> What were the problems with the chemically etched compared to laser
> cut? IOW, please explain "work anywhere near as well"?
>


My guess would be the etched edges are nowhere near clean, they will 
release
the paste harder i reckon.

ST

> >> In the past, one could get chemically etched stencils, but they did
> >> not work anywhere near as well as the laser cut ones.
> >
> > What were the problems with the chemically etched compared to laser
> > cut? IOW, please explain "work anywhere near as well"?

The chemical etching was difficult to control to keep the sidewalls of the
etched area flat and uniform. The variability of the process was a real
problem.

We might have one stencil that was perfect, others would create multiple
problems with paste release. There were also headaches with the etch being
non-uniform, thus one often times had to tweak the paste mask and change
sizes depending on pad size and aspect ratio.

Here is a link to a fellow that gives an overview of both processes.
http://www.smt.com/p8.htm

Thanks
Ron

Hi all,

following the discussion about the solder paste stencils i wonder
if it is possible to dispense the solder paste with a cnc machine.

In my opinion a stencil does not make much sense if you only have one pcb,
or maybe very few.

Would if work to mount a syringe type dispenser to a cnc?
I think it might be a big problem to ensure the paste bead is really placed
on the board and doesn't stick to the tip and is smeared all over the 
place.


Just wondering if it would be possible with no intention to build.

ST

Stefan,

Yes. Not only possible, but easy.

Ballendo

--- In Homebrew_PCBs@yahoogroups.com, Stefan Trethan 
<stefan_trethan@g...> wrote:
> Hi all,
> 
> following the discussion about the solder paste stencils i wonder
> if it is possible to dispense the solder paste with a cnc machine.
> 
> In my opinion a stencil does not make much sense if you only have 
one pcb,
> or maybe very few.
> 
> Would if work to mount a syringe type dispenser to a cnc?
> I think it might be a big problem to ensure the paste bead is 
really placed
> on the board and doesn't stick to the tip and is smeared all over 
the 

> place.
> 
> 
> Just wondering if it would be possible with no intention to build.
> 
> ST

On Fri, 23 Apr 2004 11:15:00 -0000, ballendo <ballendo@...> wrote:

> Stefan,
>
> Yes. Not only possible, but easy.
>
> Ballendo

details?

ST

>> use a gerber app.  I use Pentalogix Viewmate (free, google for
>> site).  Use the CAM processor in Eagle to output gerber rs247x
>> files.  You can import multiple gerbers into viewmate and combine
>> them to make a panel.
>
>   Phil, does the drill size are imported ok ? For TT transfer.
>
> thx,
> Vasile
>


If you are using TT you most likely don't need the gerber files?

Why don't you panelize with a picture editing software?


ST

On Thu, 22 Apr 2004, Phil wrote:

> use a gerber app.  I use Pentalogix Viewmate (free, google for
> site).  Use the CAM processor in Eagle to output gerber rs247x
> files.  You can import multiple gerbers into viewmate and combine
> them to make a panel.

  Phil, does the drill size are imported ok ? For TT transfer.

thx,
Vasile

>   because usualy is altering the scale factor...
>

not necessarily.

e.g. print to pdf and export with acrobat to a pixel image.
then you get 1:1. of course you can also use post script or something.
Maybe your layout software allows to export to images directly - then 
select
the resolution of the printer you are going to use.

ST

On Fri, 23 Apr 2004, Stefan Trethan wrote:

>
> >> use a gerber app.  I use Pentalogix Viewmate (free, google for
> >> site).  Use the CAM processor in Eagle to output gerber rs247x
> >> files.  You can import multiple gerbers into viewmate and combine
> >> them to make a panel.
> >
> >   Phil, does the drill size are imported ok ? For TT transfer.
> >
> > thx,
> > Vasile
> >
>
>
> If you are using TT you most likely don't need the gerber files?
>
> Why don't you panelize with a picture editing software?
>
  because usualy is altering the scale factor...

I've not been able to figure out how to get the drill holes but its 
not really a problem for me.  I just drill in the center of the pad.  

You can output the drill layer in eagle but its symbols, not actual 
holes.  I've not really looked into how to get the holes, though so 
it might be possible.


--- In Homebrew_PCBs@yahoogroups.com, Vasile Surducan <vasile@s...> 
wrote:

> 
> 
> 
> 
> On Thu, 22 Apr 2004, Phil wrote:
> 
> > use a gerber app.  I use Pentalogix Viewmate (free, google for
> > site).  Use the CAM processor in Eagle to output gerber rs247x
> > files.  You can import multiple gerbers into viewmate and combine
> > them to make a panel.
> 
>   Phil, does the drill size are imported ok ? For TT transfer.
> 
> thx,
> Vasile

--- In Homebrew_PCBs@yahoogroups.com, Vasile Surducan <vasile@s...> 
wrote:
> 
> 
> 
> 
> On Fri, 23 Apr 2004, Stefan Trethan wrote:
> 
> >
> > >> use a gerber app.  I use Pentalogix Viewmate (free, google for
> > >> site).  Use the CAM processor in Eagle to output gerber rs247x
> > >> files.  You can import multiple gerbers into viewmate and 
combine
> > >> them to make a panel.
> > >
> > >   Phil, does the drill size are imported ok ? For TT transfer.
> > >
> > > thx,
> > > Vasile
> > >
> >
> >
> > If you are using TT you most likely don't need the gerber files?
> >
> > Why don't you panelize with a picture editing software?
> >
>   because usualy is altering the scale factor...

and when doing panels of double sided boards, you want very precise 
placement for both sides.  plus, the pentalogix product is well 
designed for precisely this application and its free.

>
> and when doing panels of double sided boards, you want very precise
> placement for both sides.  plus, the pentalogix product is well
> designed for precisely this application and its free.
>
>

and i encourage using the pentalogix product.
But it is not like it would be that hard to do with a image editing 
software.
just enlarge the left border by the width, and paste the image in the 
space.
this results in the placement on the exact pixel.

Of course it is not the ideal solution - but if anything else fails it is 
not like
there is no other way.
There is also picture editing software which can very well cope with 
multiple layers
in one picture.


ST

On Fri, 23 Apr 2004 18:15:41 -0000, Phil <phil1960us@...> wrote:

> I've not been able to figure out how to get the drill holes but its
> not really a problem for me.  I just drill in the center of the pad.
>
> You can output the drill layer in eagle but its symbols, not actual
> holes.  I've not really looked into how to get the holes, though so
> it might be possible.
>

I thought that too.
but my layout software did the pad in one go, and then did a "negative" 
hole.
of course the plotter has no negative color. when printing with the laser
it works fine...

Drilling in the center works, but it is easier with the guiding hole for 
me.

It looked on the images like there is a short with pad and track between 
pads?

ST

Stefan Trethan wrote:
> 
> 
> no, i get copper with no hydrogen too.
> I aim for a area ratio that starts to produce hydrogen and chlorine at the 
> same time,
> and then use current slightly below that point.
> It is better if i get hydrogen before chlorine because that means the 
> copper area
> is the absolute minimum. if i get clorine earlier then the copper could 
> still be made
> smaller.

Its not good idea to have cathode CD (current density) at level where 
hydrogen is almost being produced. I would operator at about half the CD of 
the CD needed to produced hydrogen. The ratio of grams of recovered copper 
per unit charge depends on the CD and solution agitation near cathode 
surface. You trying to maximize this ratio.

Here is a copy of a document I wrote when doing experiments with alkaline 
ammonia sulfate electrolytic regeneration.

http://members.optusnet.com.au/~eseychell/etching_eff.html


> 
> Yesterday i let it run for several hours with some 200mA and at then end i 
> got visible regeneration.
> I didn't want to run it overnight without supervision so i turned it off 
> and pulled ou the copper.
> now i will go and switch it on again i think.
> 
> The lid for the vertical tank seems to be set now, but i have still some 
> work to do before i can
> activate it.
> It will hold 1 liter of etchant, and accept boards up to 35cm x 20cm which 
> is more than i need.
> It is also about the size i can make a TT (A4).
> I'm stuck right now with creating the lid, it got more complicated since i 
> aimed for electrolytic
> regeneration. Right now i think i will seperate the lid in 3 sections, the 
> middle conaining
> the pcb clamp and the outer 2 containing the electrolysis electrodes. This 
> way i can input/output
> boards without disturbing electrolysis and i can also completely remove 
> the electrodes if i should
> once need the whole 35cm length. I will most likely make the copper 
> electrode holding lid twice,
> one holding a wire, the other blank, because obviously the copper needs to 
> b taken out
> if there is no current present.
> 
> 
> I would have one other question now - do you think the etchant penetrates 
> the welding rods?

the graphite rods are a somewhat porous, I guess a millimeter or two of 
penetration.

> I connected them by drilling a hole and inserting a wire. then i sealed it 
> with silicone.
> If the etchant seeps through this is no good of course.

Epoxy would be best choice for a wire sealant. Heat the rod red hot near 
where you want to connect the wire. This will remove all traces of etchant 
and copper salts absorbed in the graphite.

> Its not good idea to have cathode CD (current density) at level where
> hydrogen is almost being produced. I would operator at about half the CD 
> of
> the CD needed to produced hydrogen. The ratio of grams of recovered 
> copper
> per unit charge depends on the CD and solution agitation near cathode
> surface. You trying to maximize this ratio.
>
> Here is a copy of a document I wrote when doing experiments with alkaline
> ammonia sulfate electrolytic regeneration.
>
> http://members.optusnet.com.au/~eseychell/etching_eff.html
>
>


I thought i must decrease the cathode area to absolute minimum because 
otherwise
the etch rate might overcome the plating rate.

I will read the page...

thanks

ST

--- In Homebrew_PCBs@yahoogroups.com, Stefan Trethan 
<stefan_trethan@g...> wrote:
> On Fri, 23 Apr 2004 11:15:00 -0000, ballendo <ballendo@y...> wrote:
> 
> > Stefan,
> >
> > Yes. Not only possible, but easy.
> >
> > Ballendo
> 
> details?
> 
> ST


The syringe can be 'fed' with a stepper.  the syringe can be moved 
just like a spindle on the X/Y/Z so it would be an extreemly simple 
matter to move it to a fraction of a mm from the board, and then 
dispense any quantity.

I imagine that a pair of drops on a pad would be as easy as moving to 
the X/Y, then to Z, dispense while retracting slowly.
mmove to next X/Y and repeat.  

This would be very very simple to do both from the dispensing side of 
the machine as well as the X/Y locations.

Since this would be a non-contact device, it could easily be clamped 
on the side of your existing head and by only adding one additional 
stepper for dispensing, you would have a complete unit.

Of course, my question is where to get low cost solder paste ?

Dave




Dave

--- In Homebrew_PCBs@yahoogroups.com, Stefan Trethan 
<stefan_trethan@g...> wrote:
> Hi all,
> 
> following the discussion about the solder paste stencils i wonder
> if it is possible to dispense the solder paste with a cnc machine.
> 
> In my opinion a stencil does not make much sense if you only have 
one pcb,
> or maybe very few.

I looked into stencils and though about the use of a stainless 
stencil the size of a credit card or similar.  then one could have 
one for 8pin SOIC, 28 pin SOIC and so on.  Often there is room to do 
a few parts and the rest is by hand.

Dave

> Of course, my question is where to get low cost solder paste ?
>
> Dave
>

nowhere. i have to pay eur10/syringe.


i wonder if the paste separates from the nozzle so uniformly that you get 
the exact
same amount on each pad.

ST

> I looked into stencils and though about the use of a stainless
> stencil the size of a credit card or similar.  then one could have
> one for 8pin SOIC, 28 pin SOIC and so on.  Often there is room to do
> a few parts and the rest is by hand.
>
> Dave
>

how do you apply the paste into the openings of the stencil?
some kind of squeegee action?


ST

--- In Homebrew_PCBs@yahoogroups.com, Stefan Trethan 
<stefan_trethan@g...> wrote:
> 
> > Of course, my question is where to get low cost solder paste ?
> >
> > Dave
> >
> 
> nowhere. i have to pay eur10/syringe.
> 
> 
> i wonder if the paste separates from the nozzle so uniformly that 
you get 
> the exact
> same amount on each pad.
> 
> ST


This is one of those things you find out about machine control.

with some practice, you find that by being 0.5mm from the pcb, and 
starting the paste to flow and presising the syringe 0.1mm, that some 
exact amount is dispensed.

this will be repeated on every occurance.

One thing that comes to mind is the surface adhesion.  If the paste 
does not adhere to the surface, it will not stick and the next pad 
will get double.

But, each of these things are things you can control.

Dave

--- In Homebrew_PCBs@yahoogroups.com, Stefan Trethan 
<stefan_trethan@g...> wrote:
> 
> > I looked into stencils and though about the use of a stainless
> > stencil the size of a credit card or similar.  then one could have
> > one for 8pin SOIC, 28 pin SOIC and so on.  Often there is room to 
do
> > a few parts and the rest is by hand.
> >
> > Dave
> >
> 
> how do you apply the paste into the openings of the stencil?
> some kind of squeegee action?
> 
> 
> ST

yup.

http://www.stencilsunlimited.com/

>
>
> This is one of those things you find out about machine control.
>
> with some practice, you find that by being 0.5mm from the pcb, and
> starting the paste to flow and presising the syringe 0.1mm, that some
> exact amount is dispensed.
>
> this will be repeated on every occurance.
>
> One thing that comes to mind is the surface adhesion.  If the paste
> does not adhere to the surface, it will not stick and the next pad
> will get double.
>
> But, each of these things are things you can control.
>
> Dave
>
>

You basically say whatever the machine does wrong, it does it wrong the 
same way every time, right?

What about the software, how do i get the file i need to feed the cnc?
No point in building a expensive cnc machine and then it is impossible to 
get the right file.

I reckon the "generating the file" thing would be even harder if one 
thinks of pick and place?


ST

--- In Homebrew_PCBs@yahoogroups.com, Stefan Trethan 
<stefan_trethan@g...> wrote:
> 
> >
> >
> > This is one of those things you find out about machine control.
> >
> > with some practice, you find that by being 0.5mm from the pcb, and
> > starting the paste to flow and presising the syringe 0.1mm, that 
some
> > exact amount is dispensed.
> >
> > this will be repeated on every occurance.
> >
> > One thing that comes to mind is the surface adhesion.  If the 
paste
> > does not adhere to the surface, it will not stick and the next pad
> > will get double.
> >
> > But, each of these things are things you can control.
> >
> > Dave
> >
> >
> 
> You basically say whatever the machine does wrong, it does it wrong 
the 
> same way every time, right?
> 
> What about the software, how do i get the file i need to feed the 
cnc?
> No point in building a expensive cnc machine and then it is 
impossible to 
> get the right file.
> 
> I reckon the "generating the file" thing would be even harder if 
one 
> thinks of pick and place?
> 
> 
> ST


I think the software would be pretty easy if you can write software.

TurboCNC is a very low cost package that handles g-code.

If you call out a drill hole for every SMT pad, or wherever you want 
to dispense paste, then the g-code can use that as the basis for 
moving.  You have full control over Z axis speed rate, or feed rate.

so you can rapid to a point and then the driller would think it was 
going to drill, but your special code would tell it to dispense.

I would do that in Excel only because I am do familiar with Excel and 
have no problem with G-code.

Dave

>
> I think the software would be pretty easy if you can write software.
>
> TurboCNC is a very low cost package that handles g-code.
>
> If you call out a drill hole for every SMT pad, or wherever you want
> to dispense paste, then the g-code can use that as the basis for
> moving.  You have full control over Z axis speed rate, or feed rate.
>
> so you can rapid to a point and then the driller would think it was
> going to drill, but your special code would tell it to dispense.
>
> I would do that in Excel only because I am do familiar with Excel and
> have no problem with G-code.
>
> Dave


I expected that...
I have a seperate "solder paste" layer which carries apertures where the 
pads are i think.
of course i also could modify the library to make another layer.

What i wonder is not so much how to find the right position, but how to 
find the right amount.
How can i influence the time the dispenser is on (or the mm the screw is 
operated)?

I basically have something like the drill file, that has every position of 
a pad stored,
so the dispenser is already moving and positioned over the pad "like if it 
would drill".
how is now the spindle lowered? (or in this case lowered and the paste 
pressed out)
where and when in the software do you set things like spindle feedrate, 
and depth?


thanks

ST

Stefan,

If your cnc machine was set up as Dave suggests, you'd use a couple 
of its "normal" features to make this easy:

1)Incremental motion, which is G91 (considers previous/starting point 
as zero each time)

2)Canned cycles for "drilling" G81 (this makes your cnc file just a 
set of locations, as each time the G81 is called, a series of moves--
designed for drilling-- are made. We wouldn't be drilling, but with 
some creative use of the canned cycle parameters, we would achieve 
the desired motion--as Dave already described. 

If you add the ability to use Macros--which my cheap cnc machines DO 
have--then it gets even easier. A macro is just a set of instructions 
that YOU (or anyone) creates and holds "together" under one code, 
like M345.

Now the dispensing can be done using either of these methods; you 
will just have more control if Macros can be created and used. Not 
all cnc machines have them...

So you would set up the canned cycle drill params for different 
sized "dots"--same with macros--. The difference is that with macros 
you'd have a DIFFERENT macro for each line, dot or even entire SMD 
package that you want to "lay down" paste for. With the drill cycle
(s), you'd have to change the parameters for each set of dots of a 
given size. Then do the next size, and so on...

Dave left out one critical part, which is that the stepper driving 
the paste plunger needs to reverse a bit after each dot is placed. 
This keeps a "string" from hanging between dot placements. (Any one 
who's ever used a caulking gun knows what I'm talking about<G>) The 
drill cycle "retract" does this for us.

Hope this helps,

Ballendo

PS. You will actually be using TWO drill cycles per dot. One for the 
paste syringe, and one for the z axis movement. With macros these two 
can be combined into ONE macro call. And the rest of the file, as 
written above, is just the xy coords of your dots. (The canned cycles 
are "persistant" until something else is called, so the desired 
action takes place at each defined XY location.) My machines will do 
this.

PPS. For this "screw drive" method of dispensing to work well, you 
need a stiff syringe, and a SOLID mount for it. Because "flex" plays 
havoc with dot size, as the syringe pluger moves downward. At the 
top, the entire syringe tube can flex to absorb some of the downward 
movement without dispensing; so as you get lower the dot size can 
increase. Fortunately we have another couple "normal" cnc functions 
which we can use to deal with this, which are tool offsets, and tool 
wear offsets. Creative use of these will acount for the problem 
mentioned.  
 
And the "other" choice is to use a dispenser from (or modeled after) 
those industrial units sold by EFD, and others...








--- In Homebrew_PCBs@yahoogroups.com, Stefan Trethan 
<stefan_trethan@g...> wrote:
> 
> >
> >
> > This is one of those things you find out about machine control.
> >
> > with some practice, you find that by being 0.5mm from the pcb, and
> > starting the paste to flow and presising the syringe 0.1mm, that 
some
> > exact amount is dispensed.
> >
> > this will be repeated on every occurance.
> >
> > One thing that comes to mind is the surface adhesion.  If the 
paste
> > does not adhere to the surface, it will not stick and the next pad
> > will get double.
> >
> > But, each of these things are things you can control.
> >
> > Dave
> >
> >
> 
> You basically say whatever the machine does wrong, it does it wrong 
the 
> same way every time, right?
> 
> What about the software, how do i get the file i need to feed the 
cnc?
> No point in building a expensive cnc machine and then it is 
impossible to 
> get the right file.
> 
> I reckon the "generating the file" thing would be even harder if 
one 

> thinks of pick and place?
> 
> 
> ST

Hello,

Most pcb layout software of decent quality will output drill files, 
which as just described, can be easily modified to become dispensing 
files. And there are other ways as well.

And the same can be said for the picn and place aspects. Expensive 
ecad will have these sorts of things inherent, and less expensive 
types(like Eagle can use ULP's) usually have a mthod to create these 
types of files. 

Face it, what we're discussing is not brand new to the pcb fab world!

Ballendo

PS. I bet you've already started a reply which says, "But with SMD's 
you DON't drill holes..." Gotta think outside the box sometimes, but 
the capability is there, and easily obtained.

Quite a lot of answering for something you "won't be doing" anyway<G> 

> --- In Homebrew_PCBs@yahoogroups.com, Stefan Trethan 
> <stefan_trethan@g...> wrote:
> > What about the software, how do i get the file i need to feed the 
> cnc?
> > No point in building a expensive cnc machine and then it is 
> impossible to 
> > get the right file.
> > 
> > I reckon the "generating the file" thing would be even harder if 
> one 
> > thinks of pick and place?
> > 
> > 
> > ST
>

Hello Stefan,

This is done by empirical testing/experimentation at first, using the 
canned cycle and/or macro params. Then it will be repeatable.

I already explained the screw. With the air pressure syringe types, 
the timing  can be controlled using G04, which is a dwell in 
milliseconds. 

Hope this helps,

Ballendo

P.S. I have never used an excel program to create a toolpath. Others 
have had great success doing so; I've just never found it necessary.

--- In Homebrew_PCBs@yahoogroups.com, Stefan Trethan 
<stefan_trethan@g...> wrote:
> 
> What i wonder is not so much how to find the right position, but 
how to 
> find the right amount.
> How can i influence the time the dispenser is on (or the mm the 
screw is 
> operated)?

Thanks Ballendo,

I understand now that my problem definitely will be the software, not the 
hardware..
I guess you meant it to sound easy but it is all a bit intimidating.

I want this thing as a help, for making faster and more precise smd boards.
If the building (and especially software stuff) gets more complicated and 
time demanding
than applying the paste by hand it is not really a good idea.

I'm not a software guy. I hate fiddling with software to get things going. 
I'd rater
apply solder paste 10 hours by hand than trying to get software to work 
for 10 hours.

How long would you estimate it takes to get all the software going?

How long would it take to get the drilling software running if my layout 
software produces
excellon drill files?
Is this possible with free software only?

How long would it take to get the software running, and roughly adjusted 
for solder paste
dispensing.
Assume a rx274 file as a output with different size "dots" for different 
paste amounts in the
right places.

I know this is nearly impossible to estimate but i need to know if it is 
like
"oh yes, that would be simple, i know how i would solve that" or "that's 
tough, nobody
has tried this before".

Which software would i need for solder paste dispensing, is it possible 
with free software?

If possible i would like to download software first, have a look at it 
etc. I don't want
to build the machine and then the whole thing fails because of software 
problems.


thanks a lot for the other hints, i agree on the mechanical things.

The software thing is really intimidating for me, the question is not if i 
can solve
such problems but more if i want to spend a lot of time with it.


ST

On Sun, 25 Apr 2004 11:07:19 -0000, ballendo <ballendo@...> wrote:

> Stefan,
>
> If your cnc machine was set up as Dave suggests, you'd use a couple
> of its "normal" features to make this easy:

On Sun, 25 Apr 2004 11:15:32 -0000, ballendo <ballendo@...> wrote:

> Hello,
>
> Most pcb layout software of decent quality will output drill files,
> which as just described, can be easily modified to become dispensing
> files. And there are other ways as well.

easily modified - how does one do that.
that is the point i don't understand, is there some software which will do 
that
for me or do i have to edit the file per hand, taking as much time as
just dispensing the paste by hand?

>
> And the same can be said for the picn and place aspects. Expensive
> ecad will have these sorts of things inherent, and less expensive
> types(like Eagle can use ULP's) usually have a mthod to create these
> types of files.
>
> Face it, what we're discussing is not brand new to the pcb fab world!

It is brand new to me ;-)
And maybe it is not that common to homebrew (couldn't find a single page 
on the web).
I really have no idea how one could possibly generate a file for pick and 
place,
but it is no priority, only drill and paste would a lot of help already.

>
> Ballendo
>
> PS. I bet you've already started a reply which says, "But with SMD's
> you DON't drill holes..." Gotta think outside the box sometimes, but
> the capability is there, and easily obtained.

No, i have another reply started (and sent) already.
i completely understand the "fake drill is paste dot" part.
I reckon i would make 2 "drill" layers, or use the solder paste layer like 
a drill layer,
to allow for mixed design.

>
> Quite a lot of answering for something you "won't be doing" anyway<G>
>
I won't be doing it JUST NOW.. i want to know if i have to look out for 
proper slides
and stuff, i have it in the back of my head for some time already...
There are other more important things that need time & money now, but i 
want
to find out if i need to set the cnc on the list or not.
And i fear this demands for finding out how i would do it.

thanks

ST

On Sun, 25 Apr 2004 11:21:52 -0000, ballendo <ballendo@...> wrote:

> Hello Stefan,
>
> This is done by empirical testing/experimentation at first, using the
> canned cycle and/or macro params. Then it will be repeatable.
>
> I already explained the screw. With the air pressure syringe types,
> the timing  can be controlled using G04, which is a dwell in
> milliseconds.
>
> Hope this helps,
>
> Ballendo
>
> P.S. I have never used an excel program to create a toolpath. Others
> have had great success doing so; I've just never found it necessary.
>

as in my other post, how do you insert these gerber commands?
Do you edit the file all by hand or do you have a software where you just 
load the file?

ST

it's usually the CAM software not the hardware when it comes to macros, I
think one would need 4 axis hardware though, with the forth being the
specialty drive for the dispenser, you could use G81 to position, and place
a M code between each G81 entry, have the M call the 4th axis movement sub
routine. You use the Z and R variables to control depth of plunge, and safe
height with G81.
This could all be done easily and freely using Eagle lite, and Mach2 demo
(although I highly advise actually buying the software in the long run- then
you can actually have enough room to mill stuff, but the demo allows plenty
of room for canned cycles. , at least for the boards that fit in eagle lite
: )

I really think this topic probably would be better suited to the cnc group,
but since the people involved with this subject frequent the same places,
the topics blur the lines.


JT


----- Original Message ----- 

From: "ballendo" <ballendo@...>
To: <Homebrew_PCBs@yahoogroups.com>
Sent: Sunday, April 25, 2004 7:07 AM
Subject: [Homebrew_PCBs] Re: solder paste - cnc dispenser


> Stefan,
>
> If your cnc machine was set up as Dave suggests, you'd use a couple
> of its "normal" features to make this easy:
>
> 1)Incremental motion, which is G91 (considers previous/starting point
> as zero each time)
>
> 2)Canned cycles for "drilling" G81 (this makes your cnc file just a
> set of locations, as each time the G81 is called, a series of moves--
> designed for drilling-- are made. We wouldn't be drilling, but with
> some creative use of the canned cycle parameters, we would achieve
> the desired motion--as Dave already described.
>
> If you add the ability to use Macros--which my cheap cnc machines DO
> have--then it gets even easier. A macro is just a set of instructions
> that YOU (or anyone) creates and holds "together" under one code,
> like M345.
>
> Now the dispensing can be done using either of these methods; you
> will just have more control if Macros can be created and used. Not
> all cnc machines have them...
>
> So you would set up the canned cycle drill params for different
> sized "dots"--same with macros--. The difference is that with macros
> you'd have a DIFFERENT macro for each line, dot or even entire SMD
> package that you want to "lay down" paste for. With the drill cycle
> (s), you'd have to change the parameters for each set of dots of a
> given size. Then do the next size, and so on...
>
> Dave left out one critical part, which is that the stepper driving
> the paste plunger needs to reverse a bit after each dot is placed.
> This keeps a "string" from hanging between dot placements. (Any one
> who's ever used a caulking gun knows what I'm talking about<G>) The
> drill cycle "retract" does this for us.
>
> Hope this helps,
>
> Ballendo
>
> PS. You will actually be using TWO drill cycles per dot. One for the
> paste syringe, and one for the z axis movement. With macros these two
> can be combined into ONE macro call. And the rest of the file, as
> written above, is just the xy coords of your dots. (The canned cycles
> are "persistant" until something else is called, so the desired
> action takes place at each defined XY location.) My machines will do
> this.
>
> PPS. For this "screw drive" method of dispensing to work well, you
> need a stiff syringe, and a SOLID mount for it. Because "flex" plays
> havoc with dot size, as the syringe pluger moves downward. At the
> top, the entire syringe tube can flex to absorb some of the downward
> movement without dispensing; so as you get lower the dot size can
> increase. Fortunately we have another couple "normal" cnc functions
> which we can use to deal with this, which are tool offsets, and tool
> wear offsets. Creative use of these will acount for the problem
> mentioned.
>
> And the "other" choice is to use a dispenser from (or modeled after)
> those industrial units sold by EFD, and others...
>
>
>
>
>
>
>
>
> --- In Homebrew_PCBs@yahoogroups.com, Stefan Trethan
> <stefan_trethan@g...> wrote:
> >
> > >
> > >
> > > This is one of those things you find out about machine control.
> > >
> > > with some practice, you find that by being 0.5mm from the pcb, and
> > > starting the paste to flow and presising the syringe 0.1mm, that
> some
> > > exact amount is dispensed.
> > >
> > > this will be repeated on every occurance.
> > >
> > > One thing that comes to mind is the surface adhesion.  If the
> paste
> > > does not adhere to the surface, it will not stick and the next pad
> > > will get double.
> > >
> > > But, each of these things are things you can control.
> > >
> > > Dave
> > >
> > >
> >
> > You basically say whatever the machine does wrong, it does it wrong
> the
> > same way every time, right?
> >
> > What about the software, how do i get the file i need to feed the
> cnc?
> > No point in building a expensive cnc machine and then it is
> impossible to
> > get the right file.
> >
> > I reckon the "generating the file" thing would be even harder if
> one
> > thinks of pick and place?
> >
> >
> > ST
>
>
>
>
> Be sure to visit the group home and check for new Bookmarks and files:
> http://groups.yahoo.com/group/Homebrew_PCBs
> Yahoo! Groups Links
>
>
>
>
>

--- In Homebrew_PCBs@yahoogroups.com, Stefan Trethan 
<stefan_trethan@g...> wrote:
> On Sun, 25 Apr 2004 11:21:52 -0000, ballendo <ballendo@y...> wrote:
> 
> > Hello Stefan,
> >
> > This is done by empirical testing/experimentation at first, using 
the
> > canned cycle and/or macro params. Then it will be repeatable.
> >
> > I already explained the screw. With the air pressure syringe 
types,
> > the timing  can be controlled using G04, which is a dwell in
> > milliseconds.
> >
> > Hope this helps,
> >
> > Ballendo
> >
> > P.S. I have never used an excel program to create a toolpath. 
Others
> > have had great success doing so; I've just never found it 
necessary.
> >
> 
> as in my other post, how do you insert these gerber commands?
> Do you edit the file all by hand or do you have a software where 
you just 
> load the file?
> 
> ST


I use TurboCNC and in that they have a jump command,  think of a goto 
in prgramming.

I move to the X/Y location, then jump to the drilling routine.

The drilling routine tells the Z to feed, and includes feed rate and 
feed depth.  it also included retraction speed and retraction height.

It would be a simple matter to add a second axis in that routing to 
dispense some value of paste.

I immeadiatly think that one could create some common units of paste.

A 0603 cap might get 6 units per pad, an 0805 might get 8 units.  A 
SOIC chip might get 6 units per pad, a TSSOP might get 3 units and so 
on.

With some programming, this should be simple enough and I am sure 
that in the near future, someone will offer the software for us home-
brew types.

Dave

>
>
> I use TurboCNC and in that they have a jump command,  think of a goto
> in prgramming.
>
> I move to the X/Y location, then jump to the drilling routine.

  So you actually manually edit the code, by hand, on a command to command 
basis?

ST

> I'm not a software guy. I hate fiddling with software to get things 
going. 
> I'd rater
> apply solder paste 10 hours by hand than trying to get software to 
work 
> for 10 hours.
> 
> How long would you estimate it 
> takes to get all the software going?
> 
> How long would it take to get the 
> drilling software running if my 
> layout software produces excellon drill files?


For an fast programmer and a fixed file type and a fixed output code, 
this could be pretty fast.  couple of days to a couple of weeks.


> Is this possible with free software only?
> 
> How long would it take to get the 
> software running, and roughly adjusted 
> for solder paste dispensing ?
> Assume a rx274 file as a output 
> with different size "dots" for different 
> paste amounts in the
> right places.

To adjust the software, I think it would be a matter of a few boards, 
or a dozen pads.

The key is to dispense the EXACT same amount from a new tube as from 
the middle and the end.

 
> I know this is nearly impossible 
> to estimate but i need to know if it is 
> like "oh yes, that would be simple, 
> I know how I would solve that" or "that's 
> tough, nobody
> has tried this before".
> 
> Which software would i need for solder 
> paste dispensing, is it possible 
> with free software?
> 
> If possible i would like to download 
> software first, have a look at it 
> etc. I don't want to build the machine 
> and then the whole thing fails because of 
> software problems.


AS far as I know, it is not currently available from free software 
for the whole process.

As I see the process, it would be to take the gerber paste file and 
possibly the NCDrill file and merge them.

Then, take that output and generate another file for paste dispensing.

I see the simplicity of drops on specific locations, but it may be 
that a line paste, like how an inkjet works, might be better.  the Z 
would drop as the X line was being moved, and at the drop of the Z, a 
constant dispensing occured. then rate of movement would determine 
quantity of dispensing.


My bottom line on the idea is to build the CNC driller.  that machine 
will work now, for software that is available now.  the paste 
dispenser will be an add-on for that by the time you are done 
building.

And while building, start collecting the parts for the dispenser add-
on.

Who knows, Ballendo may offer this as an option on his machines 
pretty soon !

Dave

--- In Homebrew_PCBs@yahoogroups.com, Stefan Trethan 
<stefan_trethan@g...> wrote:
> >
> >
> > I use TurboCNC and in that they have a jump command,  think of a 
goto
> > in prgramming.
> >
> > I move to the X/Y location, then jump to the drilling routine.
> 
>   So you actually manually edit the code, by hand, on a command to 
command 
> basis?
> 
> ST


I have the Excel automated to do that.

If I ever get around to learning Visual Basic I'll put it all into a 
complete software package.

But, the drilling I am talking about is not for PCB's but for some 
other equipment.  I have to write the whole G-code for the hole 
locations and that comes from some mathematical models.

Dave

thanks Dave.....

i don't like the "couple of weeks" part,
an the "fast programmer".

i could of course edit the librarys so that there is a short line on the 
place
where the paste has to go. one could use multiple lines, and vary the 
lengt to change amount.

I still don't get the part where it moved from a 2d excellon drill file or 
gerber file
to a 3d cnc file, where and how do i add the drill depth, feed speed, etc?

Is there a really simple solution to modify these "z axis" action to get a 
dispenser
to "mill the paste above the pad"?
I mean imagine you have a 2d drawing of a part you want to mill, how do 
you add the
z axis? how do you tell how fast the dispenser goes down, and how far it 
moves?

Maybe the "pump paste now" could be hardware engaged - whenever the nozzle 
is down fire up..
one could use adjustable delays and "suck back" as soon as the nozzle is 
raised.

ST

--- In Homebrew_PCBs@yahoogroups.com, "Jeremy Taylor" <jt@j...> wrote:
> it's usually the CAM software not the hardware when it comes to 
macros, I
> think one would need 4 axis hardware though, with the forth being 
the
> specialty drive for the dispenser, you could use G81 to position, 
and place
> a M code between each G81 entry, have the M call the 4th axis 
movement sub
> routine. You use the Z and R variables to control depth of plunge, 
and safe
> height with G81.
> This could all be done easily and freely using Eagle lite, and 
Mach2 demo
> (although I highly advise actually buying the software in the long 
run- then
> you can actually have enough room to mill stuff, but the demo 
allows plenty
> of room for canned cycles. , at least for the boards that fit in 
eagle lite
> : )
> 
> I really think this topic probably would be better suited to the 
cnc group,
> but since the people involved with this subject frequent the same 
places,
> the topics blur the lines.
> 
> 
> JT

I agree, but the group where this would fit well, 

http://groups.yahoo.com/group/CNC-PCB_Design/

is not a very old list and does not have the quantity of list 
members.  But the idea is that it will become the list that covers 
all the things specific to PCBs

There are lists that are more CNC oriented and I have found that many 
newbies feel overwhelmed by them.  Also, someone who wants to make a 
PCB type machine is not as interested in how to re-build a Bridgeport 
milling machine or making a plasma cutter.

But to be honest, the underlying principals both mechanical and 
software are all the same.

Dave

yea, yea, i got it...

member there some time already...

further discussion of solder paste cnc dispenser is at cnc-pcb_design


thanks

ST

--- In Homebrew_PCBs@yahoogroups.com, Stefan Trethan 
<stefan_trethan@g...> wrote:
> thanks Dave.....
> 
> i don't like the "couple of weeks" part,
> an the "fast programmer".


Compare to a newbie making a PCB with plated thru holes.....


 
> i could of course edit the librarys so that there is a short line 
on the 
> place
> where the paste has to go. 

If you can add a goto line then that is half the battle.

With the X/Y location, your table moves to those coordinates and 
stops.  The goto command would send the software to the paste 
dispensing section.


Of the top of my head, I see it as 
state a rapid to some 20 mills above the PCB.
slow feed to close to the part while also slow feeding a drop.
then slow retracting to allow the drop to remain on the board.
then a fast retract to normal Z height
(then retrun to g-code for move to next posisiton.)

Of course, if one were to have 5 drops along a SOIC pad, either the X 
or Y would have to move while dispensing, OR each of the 5 
coordinates would be needed to be a place the machine stopped and 
called the paste function.

Since there are only so many specific pad sizes, it is quite possible 
to make a routine that would so a whole 8 pin SOIC chip or a routine 
that would do specific pad sizes.

I have a variable for feed rate so that I can change that one point 
at the beginning of the program and run the part with that value.  
Basically on my machine tool, it is for drill depth as I change 
drills and find it is easier to measure the new tool and put in an 
offset than to try to get a tool to go back into exactly the same 
height as the last tool.

Dave









> Is there a really simple solution to modify these "z axis" action 
to get a 
> dispenser
> to "mill the paste above the pad"?
> I mean imagine you have a 2d drawing of a part you want to mill, 
how do 
> you add the
> z axis? how do you tell how fast the dispenser goes down, and how 
far it 
> moves?


YES.

as I laid out above, a simple goto in the existing files can send the 
software off to a sub-routine that handles the feed and dispense 
section.  That leaves your drill file very readable and puts all of 
the feed/dispense stuff into one small section.


> 
> Maybe the "pump paste now" could be hardware engaged - whenever the 
nozzle 
> is down fire up..
> one could use adjustable delays and "suck back" as soon as the 
nozzle is 
> raised.

I would look at a CNC stepper motor with a leadscrew.  imagine a 20 
turn per inch leadscrew and a 400 step per inch motor. and a quarter 
step motor driver circuit.

that is 20x400x4 or 32,000 steps per inch.  each step is 0.00003125 
inces per step.  If there are 40 mils per mm, then add a zero for 
steps per mm.

As you can see, a very tiny step will push on the syringe a very tiny 
amount.  As I said a long time ago in this thread, accuracy will go 
way up with doing CNC.

Also, if you wanted, you can get an 8 step driver that will multiply 
motor steps by 8. or a 10 step driver.  I have seen one that had up 
to 250 microsteps per motor step.

Bottom line on the dispenser it that it is the easy part.
the CNC is the easy part.

the file handling between generating your Gerber files to getting the 
paster file is (currently) the hard part.

Dave







> 
> ST

Stefan,

No. The macro or canned drill cycle parameters change(during the 
initial empirical testing with YUOR setup). 

Then these parameters are used in EACH "instance" of a dot. 

Alan's "so many units" per... is a good example. All you need is the 
program for one unit. Then you call that set of instructions whenever 
needed(at each desired XY location). I would make a macro for the 
most "typical" packages, which would include all the dot"units" for a 
given device.

To answer your question--I thought I already did with the statement 
that this is not new for pcb fab-- this stuff is NOT hard, and CAN be 
done with free software, and quickly...

For example, Mach2,TCNC, Eagle lite etc. (Have a look at Eagles ULP's 
listing. There is one already made for pick and place...)

Ballendo

P.S. I think Alan is QUITE correct when he says that "somebody"<G> 
might be providing a low cost way to do these things soon...
 

--- In Homebrew_PCBs@yahoogroups.com, Stefan Trethan 
<stefan_trethan@g...> wrote:
> >
> >
> > I use TurboCNC and in that they have a jump command,  think of a 
goto
> > in prgramming.
> >
> > I move to the X/Y location, then jump to the drilling routine.
> 
>   So you actually manually edit the code, by hand, on a command to 
command 

> basis?
> 
> ST

Hello,

Snips, inserts below...

>In Homebrew_PCBs, "Dave Mucha" <dave_mucha@y...> wrote:
ST>> How long would it take to get the 
>>drilling software running if my 
>>layout software produces excellon drill files?
>
>For an fast programmer and a fixed file type and a fixed output 
>code,this could be pretty fast.  couple of days to a couple of weeks.

I'd say a few hours; a day, Or two. MAX.

ST>> How long would it take to get the software running, and roughly 
>>adjusted for solder paste dispensing ?

>To adjust the software, I think it would be a matter of a few 
>boards, or a dozen pads.

Yes.

>The key is to dispense the EXACT same amount from a new tube as from 
>the middle and the end.

Ya think so? IMO, it is WELL proven--by the success of boards using 
hand applied "lines across tssop pads", squiggly and anything but 
exact!-- that close enough IS close enough... 

And the machine will repeat better than many humans dispensing 
manually. (This is not a slam on the machines. The human "servo" 
system is incredibly accurate; once "tuned" by practice...)

ST>>to estimate but i need to know if it is like "oh yes, that would 
>>be simple, I know how I would solve that" or "that's tough, nobody
>>has tried this before".

Yes, that would be simple. But I've been "into" CNC for decades...
 
>I see the simplicity... <Snip>then rate of movement would determine 
>quantity of dispensing.

It's still dots. Or steps of the plunger movment, if we're using the 
stepper based plunger. FWIW, the EFD dispenser designed for solder 
paste uses a leadscrew technique. BUT they put it in the housing of 
the unit, and use a flexible plunger to transmit the linear motion to 
the handpiece. like a LINEAR, instead of rotating, foredom tool...
(You can learn a lot from patents<G>)

>My bottom line on the idea is to build the CNC driller.  that 
>machine will work now, for software that is available now.  the 
>paste dispenser will be an add-on for that by the time you are done 
>building.

Yes. Or buy one from me for 500-1000 bucks... (See my next post)

>Who knows, Ballendo may offer this as an option on his machines 
>pretty soon !

Yes. Who knows...<G>

Ballendo

Hello,

Not for me. THAT group is trying to build a machine... (Poorly named 
group, IMO)

THIS group is about building PC BOARDS. 
USING machines, if doing so is possible/available/helpful. 

The same way we discuss plotter pens, laser prinintg, etchant types 
and details; we are now discussing solder paste application...

Ballendo

> Stefan Trethan wrote:
> yea, yea, i got it... member there some time already...
> further discussion of solder paste cnc dispenser is at cnc-
> pcb_design
> ST

On Fri, 23 Apr 2004, Stefan Trethan wrote:

> On Fri, 23 Apr 2004 18:15:41 -0000, Phil <phil1960us@...> wrote:
>
>
>
> Drilling in the center works, but it is easier with the guiding hole for
> me.
>
> It looked on the images like there is a short with pad and track between
> pads?
>
  A short between the track and the guiding hole on pads connected to that
track. Some does not have at all the guiding hole. But this is only with
boards imported in gerber.

best,
Vasile

I must admit I felt a bit wierd when that group was created. I thought
maybe I'd offended someone. I've never told anyone to take the
drill/mill CNC thing off the list, and I'm the list owner and -only-
moderator.

Like with the Laser Tag thing- there are a million forums and
Yahoogroups out there, one new being created every few months. And
each one has a post a week and about 3 people on each one, and they
all know each other from other Laser Tag lists.

Steve

--- In Homebrew_PCBs@yahoogroups.com, "ballendo" <ballendo@y...> wrote:

> Hello,
> 
> Not for me. THAT group is trying to build a machine... (Poorly named 
> group, IMO)
> 
> THIS group is about building PC BOARDS. 
> USING machines, if doing so is possible/available/helpful. 
> 
> The same way we discuss plotter pens, laser prinintg, etchant types 
> and details; we are now discussing solder paste application...
> 
> Ballendo
> 
> 
> 
> > Stefan Trethan wrote:
> > yea, yea, i got it... member there some time already...
> > further discussion of solder paste cnc dispenser is at cnc-
> > pcb_design
> > ST

Steve,

Thank you. It always seemed to me that drilling, paste application, 
and part placement were part of Homebrew PCBs...

Ballendo

P.S. For Mike: I don't think it's good use of group bandwidth to 
repeat the specs already given in a series of fairly recent posts. 
Please check the group archive for details, if interested. 

--- In Homebrew_PCBs@yahoogroups.com, "Steve" <alienrelics@y...> 
wrote:
> I must admit I felt a bit wierd when that group was created. I 
thought
> maybe I'd offended someone. I've never told anyone to take the
> drill/mill CNC thing off the list, and I'm the list owner and -only-
> moderator.
> 
> Like with the Laser Tag thing- there are a million forums and
> Yahoogroups out there, one new being created every few months. And
> each one has a post a week and about 3 people on each one, and they
> all know each other from other Laser Tag lists.
> 
> Steve
> 
> --- In Homebrew_PCBs@yahoogroups.com, "ballendo" <ballendo@y...> 
wrote:
> > Hello,
> > 
> > Not for me. THAT group is trying to build a machine... (Poorly 
named 
> > group, IMO)
> > 
> > THIS group is about building PC BOARDS. 
> > USING machines, if doing so is possible/available/helpful. 
> > 
> > The same way we discuss plotter pens, laser prinintg, etchant 
types 

> > and details; we are now discussing solder paste application...
> > 
> > Ballendo
> > 
> > 
> > 
> > > Stefan Trethan wrote:
> > > yea, yea, i got it... member there some time already...
> > > further discussion of solder paste cnc dispenser is at cnc-
> > > pcb_design
> > > ST

--- In Homebrew_PCBs@yahoogroups.com, "Steve" <alienrelics@y...> 
wrote:
> I must admit I felt a bit wierd when that group was created. I 
thought
> maybe I'd offended someone. I've never told anyone to take the
> drill/mill CNC thing off the list, and I'm the list owner and -only-
> moderator.
> 
> Like with the Laser Tag thing- there are a million forums and
> Yahoogroups out there, one new being created every few months. And
> each one has a post a week and about 3 people on each one, and they
> all know each other from other Laser Tag lists.
> 
> Steve

No offense to this list, but this list is how to make circuit boards, 
not how to make machines, and not how to engineer equipment, select 
linear rails or drawer slides etc.

It would be nice to have one be-all list, but this list is clearly 
stated as how to make boards.

There are other CNC lists and other machine too lists, but the 
newbies I see on those are afraid of asking questions as often the 
discussion are so high level.  if you ask about drawer slides on the 
CCED list, you will get laughed at and told 100 reasons why it will 
not work.  of couse on the CNC-PCB list you can see pictures of those 
machines and the boards they have made.

If there were not such a huge seperation between the topics, the list 
would not have been created.

Dave

--- In Homebrew_PCBs@yahoogroups.com, "ballendo" <ballendo@y...> 
wrote:
> Hello,
> 
> Not for me. THAT group is trying to build a machine... (Poorly 
named 
> group, IMO)
> 
> THIS group is about building PC BOARDS. 
> USING machines, if doing so is possible/available/helpful. 
> 
> The same way we discuss plotter pens, laser prinintg, etchant types 
> and details; we are now discussing solder paste application...
> 
> Ballendo


I have to agree that once a topic starts, and unless it would get 
more play elsewhere, it should stay on the list.

I was only agreeing with another member that the topic would fall 
more under the build a machine list.  That member recognised that 
designing a paste machine and writing software was pretty far afield 
for this list.

The thread was discussing 
- how to write the software
- how to engineer the paste mechanism
- how to write the G-code

These are pretty far off how to make a pcb as you entered assembling 
boards and machine design.  But for those who are both mechanically 
inclined and want to make a board, it is a good topic.

Pic and place would be another topic and if this list gets a lot of 
play on that topic, I would bet yet another list gets created.

> 
> 
> 
> > Stefan Trethan wrote:
> > yea, yea, i got it... member there some time already...
> > further discussion of solder paste cnc dispenser is at cnc-
> > pcb_design
> > ST

You're kidding right?

We can talk for DAYS! about CuCl regeneration (which I could give a 
rat's @$$ about), but not about engineering a way to get folks more 
productive in the "other" aspects of pcb creation?

We can talk of how to engineer a plotter into a cutter or a laser 
printer into a toner transfer device, but not how to make a usable 
paste dispenser inexpensively? 

I think that the idea of "where" a certain topic might get "more 
play" is without merit. I'm not here to get "more play". I'm here 
because I value the opinions of the folks who are HERE. You just 
posted that in the group YOU moderate, mention of a simple cheap 
machine using drawer guides would be laughed off the list! The only 
thing which seems to create "new" lists from old ones, is over 
moderation of topics...

Thankfully, that is not the case here.

Ballendo

Snips, inserts below

Perhaps we need: Homebrew PCB's-etchants, Homebrew PCB's-drilling, 
Homebrew PCB's-layout, Homebrew PCB's-phototools, Homebrew PCB's-TT, 
etc.......... NO!

>In Homebrew_PCBs, "Dave Mucha" wrote:
<snip> 
> I was only agreeing with another member that the topic would fall 
> more under the build a machine list.  That member recognised that 
> designing a paste machine and writing software was pretty far 
> afield for this list.

Reading your included snip of Stefan's reply below doesn't suggest 
that at all. It simply says there is more info AT that site ABOUT 
paste dispensing.

FWIW, A "general" cnc machine list has NO IDEA what's required FOR 
pcb's. If you're gonna make a machine for pcb's, IMO you should talk 
with the folks who are making them!
 
> The thread was discussing 
> - how to write the software
> - how to engineer the paste mechanism
> - how to write the G-code
> 
> These are pretty far off how to make a pcb as you entered 
> assembling boards and machine design.  

Dave, forgive me. But I think you've spent too long policing CCED. 
Since when does Homebrew PCB's mean only the making of the board? or 
put another way, since when does the "making" of the board stop 
before putting components on it? I think it's pretty clear that the 
most common member of homebrew'd pcb's is making pre-production 
prototypes, personal projects, and other low volume "immediate" need 
situations. With a FEW members who are in light production... For 
these folks, having methods to COMPLETE the homebrew PCB is a good 
thing...

>But for those who are both mechanically inclined and want to make a 
>board, it is a good topic.

Funny thing is, the idea is to allow someone LESS mechanically 
inclined to make (and complete) boards...
 
> Pic and place would be another topic and if this list gets a lot of 
> play on that topic, I would bet yet another list gets created.

Only if Steve wants to thin out his group. Which he doesn't appear to 
wish to do; based on his replies in this thread.

I can understand if I were promoting several thousand dollar 
solutions aimed at board productions in the hundreds. I DO have those 
too. But I'm not promoting them here, because they don't fit the 
needs of this group. A DIY drill, and a DIY solder paste dispenser, 
and a DIY pick and place DO...  

> > > Stefan Trethan wrote:
> > > yea, yea, i got it... member there some time already...
> > > further discussion of solder paste cnc dispenser is at cnc-
> > > pcb_design
> > > ST

Well said!  Jack

  ----- Original Message ----- 
  From: ballendo 
  To: Homebrew_PCBs@yahoogroups.com 
  Sent: Tuesday, April 27, 2004 1:45 PM
  Subject: [Homebrew_PCBs] Re: solder paste - cnc dispenser


  You're kidding right?

  We can talk for DAYS! about CuCl regeneration (which I could give a 
  rat's @$$ about), but not about engineering a way to get folks more 
  productive in the "other" aspects of pcb creation?

  We can talk of how to engineer a plotter into a cutter or a laser 
  printer into a toner transfer device, but not how to make a usable 
  paste dispenser inexpensively? 

  I think that the idea of "where" a certain topic might get "more 
  play" is without merit. I'm not here to get "more play". I'm here 
  because I value the opinions of the folks who are HERE. You just 
  posted that in the group YOU moderate, mention of a simple cheap 
  machine using drawer guides would be laughed off the list! The only 
  thing which seems to create "new" lists from old ones, is over 
  moderation of topics...

  Thankfully, that is not the case here.

  Ballendo

  Snips, inserts below

  Perhaps we need: Homebrew PCB's-etchants, Homebrew PCB's-drilling, 
  Homebrew PCB's-layout, Homebrew PCB's-phototools, Homebrew PCB's-TT, 
  etc.......... NO!

  >In Homebrew_PCBs, "Dave Mucha" wrote:
  <snip> 
  > I was only agreeing with another member that the topic would fall 
  > more under the build a machine list.  That member recognised that 
  > designing a paste machine and writing software was pretty far 
  > afield for this list.

  Reading your included snip of Stefan's reply below doesn't suggest 
  that at all. It simply says there is more info AT that site ABOUT 
  paste dispensing.

  FWIW, A "general" cnc machine list has NO IDEA what's required FOR 
  pcb's. If you're gonna make a machine for pcb's, IMO you should talk 
  with the folks who are making them!

  > The thread was discussing 
  > - how to write the software
  > - how to engineer the paste mechanism
  > - how to write the G-code
  > 
  > These are pretty far off how to make a pcb as you entered 
  > assembling boards and machine design.  

  Dave, forgive me. But I think you've spent too long policing CCED. 
  Since when does Homebrew PCB's mean only the making of the board? or 
  put another way, since when does the "making" of the board stop 
  before putting components on it? I think it's pretty clear that the 
  most common member of homebrew'd pcb's is making pre-production 
  prototypes, personal projects, and other low volume "immediate" need 
  situations. With a FEW members who are in light production... For 
  these folks, having methods to COMPLETE the homebrew PCB is a good 
  thing...

  >But for those who are both mechanically inclined and want to make a 
  >board, it is a good topic.

  Funny thing is, the idea is to allow someone LESS mechanically 
  inclined to make (and complete) boards...

  > Pic and place would be another topic and if this list gets a lot of 
  > play on that topic, I would bet yet another list gets created.

  Only if Steve wants to thin out his group. Which he doesn't appear to 
  wish to do; based on his replies in this thread.

  I can understand if I were promoting several thousand dollar 
  solutions aimed at board productions in the hundreds. I DO have those 
  too. But I'm not promoting them here, because they don't fit the 
  needs of this group. A DIY drill, and a DIY solder paste dispenser, 
  and a DIY pick and place DO...  

  > > > Stefan Trethan wrote:
  > > > yea, yea, i got it... member there some time already...
  > > > further discussion of solder paste cnc dispenser is at cnc-
  > > > pcb_design
  > > > ST




  Be sure to visit the group home and check for new Bookmarks and files:
  http://groups.yahoo.com/group/Homebrew_PCBs 



------------------------------------------------------------------------------
  Yahoo! Groups Links

    a.. To visit your group on the web, go to:
    http://groups.yahoo.com/group/Homebrew_PCBs/
      
    b.. To unsubscribe from this group, send an email to:
    Homebrew_PCBs-unsubscribe@yahoogroups.com
      
    c.. Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service. 




[Non-text portions of this message have been removed]

...

I personally do not care at all in which list the stuff is discussed,
as long as i am a member ;-).


I don't think that anyone tries to weaken the pcb list at all, it was more
an attempt to keep the list free of too OT things (and yes it can go OT 
with cnc.).

I also agree the other popular cnc list (edm cnc i think) is very 
intimidating, and
honestly i do not like it very much.

I don't think it is a bad idea to have the cnc list, but i also think it 
was not OT here.

Damn i get a 50 or so spam mails a day, i do not care if i have to ignore 
20 posts on a list
which have a topic that i find not interesting, and know nothing about.
That's what the subject line is for (SO PLEASE ALWAYS CHNANGE IT).

I strongly suggest to leave the decision about what it OT to the list 
owner/moderator
(who does a magnificent job if i might say so).
I also suggest that if someone mentions there might be another list which 
is even
closer to the topic that it is taken as what it was meant for, an attempt 
to help.



I still have not worked out the cnc paste despensing stuff and i suggest 
the discussion
going back there, or somewhere else. After all the CURRENT topic is 
clearly OFF TOPIC.


I still do not understand where and how those macros, or whatever they are 
called is inserted.
I mean i have a drawing of my solder paste layer, nothing more.
It contains short lines where the paste must go, that's it.
This file does know nothing about how deep the dispenser has to be 
lowered, or anything.
It is like a plotter file, just 2 dimensional.
I have zero CNC knoweledge (ok that is not exactly true) and i don't know 
how i come
 from a 2D gerber file to the signals on the pc port that control the 
motors.

Would be really, really helpful if someone could (maybe again) explain how 
this is done.
Please assume a two-and-a-half D Milling project for explanation, as it 
seems the same to me.
Assume one has a 2D gerber line-art and wants to mill it. how does he 
process the 2D file
to get the machine to move.



Thanks a lot.

ST

Just to put in my $0.02.  As a trying-to-be-a-hobbiest, I've found
this very interesting, and on topic.  It goes over things that I need
to know for building pcbs, just like the CUCL.  There are things in
this that I have not had the time/experience to know they are going to
be problems.  I don't see it as being different from discussions on
tin plating or toaster solder ovens.
  I know about ez-bake.  I also know that it doesn't have a critical
mass for getting a good discussion going.  People seem to watch it,
but there are not real on going conversations like we have here.  Now
I know about the cnc group, but seems that very smart, knowledgable
and contributing people here are leary of getting involved in that
group.  I have had no issues in homebrew PCB group with people beating
me up for stupid questions and suggestions, which is very good for an
FNG like me.
  And while I know I'm louder than a lot of people, I also know that
if I don't know stuff, then there are other people who don't, but
aren't as willing to play the fool as I am.  So as long as we don't
start talking about using quantum teleportation for parts/paste
placement,  I'm up for the discussion, and thank Steve for letting it
continue.
Richard

Stefan,

I appreciate your comments, and the posts prior to this. I've snipped 
them to get us back on track...

Okay, there's more info in those "short lines" than you think. For 
instance they have location obviously, but also a layer, and a 
filetype. And a width.

From this filetype, layer, and width info we can give our macro the 
info it needs to do the required dispensing. It is simply a matter of 
deciding WHAT a layer, or filetype, or width "means", and then USING 
that when we create the solder paste layer...

And the macro is a subroutine, so each time it's called it does the 
same thing, at the locations which are easily pulled from the drawing.

See, the thing is, you see a drawing. But the drawing is "described" 
by its file. It's that DESCRIPTION which has the "extra" info you're 
not "seeing"...

Your question about milling is answered by Eagle with the use of a 
ULP designed to create mill-able traces from the gerber files.

The SAME method can be used to make a "dispensing" ULP.

ULP is a user written program. In this case the "user" is the WRITER 
of the program; and the true end user might have NO IDEA how it's 
done; but it works just the same. No different than pressing "output 
gerber files" in your ecad once you've finished the board layout...
The conversion program uses the details provided (or "mined") to 
create the information the machine needs to mill, or drill, or 
dispense, etc.

Hope this helps,

Ballendo

P.S. Can someone please post a link to the step by step use of a ULP 
to make milling traces possible. I've only very recently begun using 
Eagle, and I have some high end--and older/outdated programs I use 
which most wouldn't be likely to use with the inexpensive machine I'm 
offering (VHF, Ztrace, and Contact2). But I'll know how to do it with 
Eagle inside and out before they ship next month<G>  FWIW, the 
excellon file is used directly for the drilling.


--- In Homebrew_PCBs@yahoogroups.com, Stefan Trethan > I mean i have 
a drawing of my solder paste layer, nothing more.
> It contains short lines where the paste must go, that's it.
> This file does know nothing about how deep the dispenser has to be 
> lowered, or anything.
> It is like a plotter file, just 2 dimensional.
<snip>
 Would be really, really helpful if someone could (maybe again) 
explain how this is done.
> Please assume a two-and-a-half D Milling project for explanation, 
as it seems the same to me.
> Assume one has a 2D gerber line-art and wants to mill it. how does 
he process the 2D file to get the machine to move.

> 
> 
> 
> Thanks a lot.
> 
> ST

--- In Homebrew_PCBs@yahoogroups.com, "ballendo" <ballendo@y...> 
wrote:
> You're kidding right?
> 
> We can talk for DAYS! about CuCl regeneration (which I could give a 
> rat's @$$ about), but not about engineering a way to get folks more 
> productive in the "other" aspects of pcb creation?

This list does not cater to what you care about, it caters to the 
topic as listed on the home page.

Read the home page of the list.  What are you reading into the 
title ?  99.99% of all PCB's are made with chemicals.  This is the 
RIGHT list for that.

> I think that the idea of "where" a certain topic might get "more 
> play" is without merit. 

If you want machine design, you don't go to electical engineers.  If 
you want to make a PCB you don't go to mechanical engineers and you 
don't go to chemical engineers to find out how to drill holes.  I 
imagine pcb design would get tossed off your list PDQ.  Why should 
all other lists but this one have a specific topic as stated on the 
home page?  Common sense should tell you where to go to get the best 
response.  Of course it starts where the idea starts.  And if your 
list started getting into the chemicals for making PCB's would you 
want to discuss that or would you tell them to come here ?  ie: go to 
where the experts are or people with similar interests.

> FWIW, A "general" cnc machine list has NO IDEA what's required FOR 
> pcb's. If you're gonna make a machine for pcb's, IMO you should 
talk 
> with the folks who are making them!

I totally disagree.  How many on your list know the chemical 
reactions of the stuff for making PCB's ?  Wrong list for that.  How 
many on here know the stress and tortional calculations for 3/4 
CRS ?  You don't ask an EE for mechanical designs and you don't as 
ME's for Chemical properties and you don't ask a chemical engineer 
how to drill holes.  (seems I already said that, but it is worth 
repeating.)

> since when does the "making" of the board stop 
> before putting components on it? 
<snip>  
> Funny thing is, the idea is to allow someone LESS mechanically 
> inclined to make (and complete) boards...

I agree, but you seem to advocating that less mechanically inclined 
design your machines.  This list is not 90% mechanical and 10% 
electronics.  just the opposite.

I think the number of posts on the topics on here reflect the 
background and experiance and direction of the list members.  90% 
want to know about or handle chemcicals.

Look at any list, you go there to get answers or offer your help.  
What part of this list description has anything to do with mechancial 
machine design ?


Dave

On Tuesday 27 April 2004 09:30 pm, Dave Mucha wrote:

> Read the home page of the list.  What are you reading into the
> title ?  99.99% of all PCB's are made with chemicals.  This is the
> RIGHT list for that.

Excluding all other methods?  That would seem to be the way it's been flowing 
lately...

<...>

> If you want machine design, you don't go to electical engineers.  If
> you want to make a PCB you don't go to mechanical engineers and you
> don't go to chemical engineers to find out how to drill holes.

I don't plan on going to *any* engineers.  The name of this list says 
"Homebrew" and I figured that maybe I'd bump into other people here who were 
making boards on a hobbyist basis...

<...>

> > FWIW, A "general" cnc machine list has NO IDEA what's required FOR
> > pcb's. If you're gonna make a machine for pcb's, IMO you should
> > talk with the folks who are making them!

> I totally disagree.  How many on your list know the chemical
> reactions of the stuff for making PCB's ?  Wrong list for that.  How
> many on here know the stress and tortional calculations for 3/4
> CRS ?  You don't ask an EE for mechanical designs and you don't as
> ME's for Chemical properties and you don't ask a chemical engineer
> how to drill holes.  (seems I already said that, but it is worth
> repeating.)

Yeah,  you already did,  so I won't repeat my response here.

<...>

> > Funny thing is, the idea is to allow someone LESS mechanically
> > inclined to make (and complete) boards...

> I agree, but you seem to advocating that less mechanically inclined
> design your machines.  This list is not 90% mechanical and 10%
> electronics.  just the opposite.

Seems like a lot of posts in here lately have been chemical,  not either one 
of the above.

> I think the number of posts on the topics on here reflect the
> background and experiance and direction of the list members.  90%
> want to know about or handle chemcicals.

Some of us don't.  I for one have absolutely *no* interest in chemical 
etching,  and plan to do it with CNC once I can get something put together. 
Is there a problem with that?

> Look at any list, you go there to get answers or offer your help.
> What part of this list description has anything to do with mechancial
> machine design ?

See above,  the last paragraph I wrote above this one.

> 
> I still have not worked out the cnc paste despensing stuff and i 
suggest 
> the discussion
> going back there, 

As I had said in one of my posts, once a topic is started and is 
getting responses, keep it on the list.  (I'm going to have to see if 
others will allow such a thing on other lists)


In Eagle, export your Gerber and NCDrill files.

open the NCDrill file and you will see pairs of numbers.  Mostly 
likely they are 6 digits and are the X and Y coordinates to your 
holes.  They are the drill hole coordinates.

Also, most likely, they are 2+4 in format, the first 2 are integer 
and the remaining 4 are decimal.

here is a small snip of one of mine.

X009200Y004950
X010400Y002250
X009400Y002250
X008400Y002250
X007400Y002250
X006400Y002250
X005400Y002250
X004400Y002250
X003400Y002250
X003400Y007050
X004400Y007050
X005400Y007050


Cartesian coordinates in X and Y that exactly line up to the drill 
holes.

OR... if you lay out pads, they could be to the pad.


You can pull this into Excel and add the G-code lines. the first line 
for example.

N10 G00 X 00.9200 Y 00.4950

N10 is just the program line number.
G00 is the G-Code for a rapid move to a location
the X and Y coordinates are that location.

This would send your machine to that location.

I then use a jump command, M60  which is very similar to a goto in 
basic.


N10 G00 X 00.9200 Y 00.4950
N15 M60    #600

#600 is line 600 in my program and where the drilling routine starts.

N600 Z -1
N610 Z 0
N620 M62

Line 600 sends the Z axis to a -1 inch. Assuming the spindle is 
always running so that would be a drill hole.

Line 610 brings the Z back to home.

Line 620 sends the program back to the END of line N15 and the 
program picks up from there to the next line.

For PASTE, we might alter this to 


N600 Z -1
N601 A -.001
N610 Z 0
N620 M62


The addition of line 601 would control the A axis (assuming they go 
XYZABC) to a -0.001 inches or mm or whatever.

That would deliver a drop of paste to the center of the pad.

IF the syringe acted like a cheap chaulk gun and kept dribbling, then 
you might use 

N600 Z -1
N601 A -.001
N602 A 0
N610 Z 0
N620 M62

Line 602 would retract the syringe back a little to remove any 
pressure from the plunger.

Excel would be very easy to use to increment the A feed and A 
retraction.

Also, one could write the code as


N600 Z -1  A -.001
N610 Z 0   A 0
N620 M62


This would typically feed BOTH the Z AND A axis so that they arrived 
at the end point exactly at the same time.  A bead would start to 
form as it was approaching the pad and be (hopefully) fully formed by 
the time it reached the pad.

there is no pause at the pad, just a reach and retract.

So, your main body would be like this

N10   X00.9200 Y00.4950
N20   M60  #100
N30   X01.0400Y00.2250
N40   M60  #100
N50   X00.9400Y00.2250
N60   M60  #100
N70   X00.8400Y00.2250
N80   M60  #100
N90   X00.7400Y00.2250
N100  M60  #100

The #100 would be the section earlier about Z movement.

The speed the unit approached the pad can be adjusted and the time on 
the pad can also be added.

All that said, this is the way I do my drilling, so if you want try 
that it is a start.

There are other ways, but all you need to do is to start on one path 
and see how it works.

Dave

> > I think the number of posts on the topics on here reflect the
> > background and experiance and direction of the list members.  90%
> > want to know about or handle chemcicals.
> 
> Some of us don't.  I for one have absolutely *no* interest in 
chemical 
> etching,  and plan to do it with CNC once I can get something put 
together. 
> Is there a problem with that?


let's try to steer this back to more PCB related questions.

There is a problem with chemical,  getting rid of the fumes and 
chemicals.

the bonus is you can do 100's of boards at the same time.

The benifit of mechanical etching is no chemicals.
the down side is the expensive cutters that wear out quickly.

A dull cutter will raise a burr on the copper and dependong on your 
hold-down method, will or will not effect the rest of the board.

It will however require intervetion to be able to solder SMT parts.

The other down side is how long it takes to make a board.  the 
mechanical engravers cut one section of a trace at a time so the 
entire time is longer than chemical etching.

Another plus for the CNC is drilled holes.

Dave



Dave

On Tuesday 27 April 2004 10:45 pm, Dave Mucha wrote:
> > > I think the number of posts on the topics on here reflect the
> > > background and experiance and direction of the list members.  90%
> > > want to know about or handle chemcicals.
> >
> > Some of us don't.  I for one have absolutely *no* interest in
> > chemical etching,  and plan to do it with CNC once I can get something put
> > together.

> > Is there a problem with that?

> let's try to steer this back to more PCB related questions.
>
> There is a problem with chemical,  getting rid of the fumes and
> chemicals.
>
> the bonus is you can do 100's of boards at the same time.

I can see where some people might end up doing that,  depending on what you 
were working on.  I see making boards at all as simply being a way to augment 
my hobby electronics stuff and make things a bit easier than the perfboard 
routine I've been using for years.

> The benifit of mechanical etching is no chemicals.
> the down side is the expensive cutters that wear out quickly.

I have the impression that this is more of a problem with some board materials 
than others,  is that the case?

> A dull cutter will raise a burr on the copper and dependong on your
> hold-down method, will or will not effect the rest of the board.

I haven't yet begun to explore the possibilities of various holddown methods. 
I can think of a few right off the top of my head,  those being simple 
clamps,  some kind of a T-slot table (more clamps),  or maybe a vacuum 
holddown setup.  What else is there?

> It will however require intervetion to be able to solder SMT parts.

That's someplace else I have no intention of going.  That stuff does provide 
some benefit,  to be sure,  in terms of component density and especially for 
manufacturing in large quantities,  but again,  I don't see me getting into 
any of that sort of thing.

> The other down side is how long it takes to make a board.  the
> mechanical engravers cut one section of a trace at a time so the
> entire time is longer than chemical etching.

True,  but it's not like I'd have to be sitting there doing it,  once I got a 
good setup going.

> Another plus for the CNC is drilled holes.

Yes,  that's a big plus in my opinion.  I'd *much* rather have the machine do 
that than have to do it myself...

> > The benifit of mechanical etching is no chemicals.
> > the down side is the expensive cutters that wear out quickly.
> 
> I have the impression that this is more of a problem with some 
board materials 
> than others,  is that the case?


Yes, FR-4 is the most common and easiest to get.  Seems no one knows 
were to get boards as the question on here rarely gets an answer.

But the phenolic boards are much gentler on the cutters, but it does 
not really matter if you can't get the darn things.


> 
> > A dull cutter will raise a burr on the copper and dependong on 
your
> > hold-down method, will or will not effect the rest of the board.
> 
> I haven't yet begun to explore the possibilities of various 
holddown methods. 
> I can think of a few right off the top of my head,  those being 
simple 
> clamps,  some kind of a T-slot table (more clamps),  or maybe a 
vacuum 
> holddown setup.  What else is there?


Duct tape.  No, really.

I have a pair of pins in the table  the board slips over them for 
alignment and from there the boards is held on the edges with duct 
tape.

The drill does not lift the board and mechanical etching is held firm 
with the two pins.




> 
> > Another plus for the CNC is drilled holes.
> 
> Yes,  that's a big plus in my opinion.  I'd *much* rather have the 
machine do 
> that than have to do it myself...


And THAT is why CNC-PCB_design was created !

the non-technical, home brew drilling concept you can build from 
stuff laying around to spending $100.00-200.00 in parts.

Dave

On Tuesday 27 April 2004 11:04 pm, Dave Mucha wrote:
> > > The benifit of mechanical etching is no chemicals.
> > > the down side is the expensive cutters that wear out quickly.
> >
> > I have the impression that this is more of a problem with some
> > board materials than others,  is that the case?

> Yes, FR-4 is the most common and easiest to get.  Seems no one knows
> were to get boards as the question on here rarely gets an answer.

That's what,  that green stuff?  Is there a FAQ about board materials 
someplace?  I saw phenolic referred to as well as FR-4,  so I figured that 
they were two different things.  <g>

> But the phenolic boards are much gentler on the cutters, but it does
> not really matter if you can't get the darn things.

They've gotta be out there somewhere.  Where *do* people go to get raw board 
material?  Only stuff I've gotten so far was an assortment pack at Radio 
Shack,  aside from some of their perfboards and such.

> > > A dull cutter will raise a burr on the copper and dependong on
> > > your hold-down method, will or will not effect the rest of the board.

> > I haven't yet begun to explore the possibilities of various
> > holddown methods.
> >
> > I can think of a few right off the top of my head,  those being
> > simple clamps,  some kind of a T-slot table (more clamps),  or maybe a
> > vacuum holddown setup.  What else is there?

> Duct tape.  No, really.

Would that be flat enough?

> I have a pair of pins in the table  the board slips over them for
> alignment and from there the boards is held on the edges with duct
> tape.

Ok.

> The drill does not lift the board and mechanical etching is held firm
> with the two pins.

I'm figuring on building a smallish table,  and then putting another hunk of 
material on top of that for "sacrificial" purposes,  so I won't have to worry 
too much about it.

> > > Another plus for the CNC is drilled holes.

> > Yes,  that's a big plus in my opinion.  I'd *much* rather have the
> > machine do that than have to do it myself...

> And THAT is why CNC-PCB_design was created !

Works for me!

> the non-technical, home brew drilling concept you can build from
> stuff laying around to spending $100.00-200.00 in parts.

And I have a *LOT* of "stuff laying around".  Got a whole box of assorted 
stepper motors scrounged out of various floppy drives and printers,  more 
transistors and chips than I want to think about,  enough passive components 
that I really don't think I'll need to buy any anytime soon,  and more.  
About the only thing I can think of that I might want that I don't already 
have is a couple of good-sized hunks of threaded rod...

--- In Homebrew_PCBs@yahoogroups.com, "Dave Mucha" <dave_mucha@y...>
wrote:
> --- In Homebrew_PCBs@yahoogroups.com, "Steve" <alienrelics@y...> 
> wrote:
> > I must admit I felt a bit wierd when that group was created. I 
> thought
> > maybe I'd offended someone. I've never told anyone to take the
> > drill/mill CNC thing off the list, and I'm the list owner and -only-
> > moderator.
...
> No offense to this list, but this list is how to make circuit boards, 
> not how to make machines, and not how to engineer equipment, select 
> linear rails or drawer slides etc.
> 
> It would be nice to have one be-all list, but this list is clearly 
> stated as how to make boards.
> 
> There are other CNC lists and other machine too lists, but the 
> newbies I see on those are afraid of asking questions as often the 
> discussion are so high level.  if you ask about drawer slides on the 
> CCED list, you will get laughed at and told 100 reasons why it will 
> not work.  of couse on the CNC-PCB list you can see pictures of those 
> machines and the boards they have made.
> 
> If there were not such a huge seperation between the topics, the list 
> would not have been created.

I understand your points, and notice I only said something when it
came up. This is "my" list only in that I started it and guide it, but
without members I'd be here by myself.

My original intention was to bring together all the separated
discussions about making PCBs that I saw on other lists. Usually
someone on one list would have the answer to someone's question on
another list, and never the twain would meet.

So if I don't mention something specifically, it does -not- mean that
it is not allowed.

An excerpt from the homepage:

********************
Making, etching, soldering printed circuit boards yourself.

Any method allowed by law and local custom including photo, CAM, toner
transfer, direct printing, plotter, laundry pen, silkscreening,
toaster oven, paint stripper heat gun, etc.
********************

So if you -want- to take some of the discussion to another list, if
you -want- to start a list dedicated to CNC_PCB, I understand as it is
a big subject. But don't feel like you -have- to.

Steve, the list owner and only moderator of Homebrew_PCBs

> So if I don't mention something specifically, it does -not- mean 
that
> it is not allowed.
> 


I know and I like the way you moderate.  I am a member of at least 
one other of your lists and the totally open method works well.

Personally, I like to keep a thread where it starts UNLESS there is a 
signifigantly better place for it.

As it appears, there are a very few on here that seem to have an 
interest is solder paste and fewwer still that feel like they could 
design such an item mechanically and I'm not sure how many could put 
the whole package together with some sort of automatic file handling 
direct from the Gerbers or NCDrill file.

I know I could whip up axis from parts I already have.  I could make 
the syringe holder and feed assembly.

I could write lots and lots of stuff in Excel to set the software, 
but unless I was doing 100 boards, I think hand paste would be faster.

I could not write a software program that took a Gerber and made G-
Code that automatically laid out the paste.

At some point we will run out of expertise on the members of this 
list.  Then for this topic to go to the next level it will be forced 
to go elswhere.  That's life.

I think Ballendo has a little desire to promote his unit and cheap 
home-builts are part of his market, or competiton.  And it looks like 
the members on here are much more intereted in buying a unit whereas 
people on CNC-PCB_Design are there to build one.  Absolutly no 
customers there at all.

I know I have had posts on his list deleted for not being on topic 
and they were really close to the topic, WAY closer than machine 
design on this list.   If this was run under rules the way he runs 
his lists, a lot of posts would have been deleted.  I think that is 
why I got my dander up a little.

Oddly CNC-PCB_Design is a 'bad' name and 'too limiting' of a list, 
yet for some reason this list is an ever expanding and encompasing 
umbrella.  just odd. (not your words)

Personally, if someone makes a dispenser, I'll be really peeved... 
if they DON'T post pictures !!!

Dave

Dave,

Here's a copy of the FIRST line ON the home page:

  Making, etching, soldering printed circuit boards yourself.

See that third word there... 

Next, The list OWNER has approved this thread. And posted as such 
already. Please respect what the LIST OWNER has said.

Ballendo

--- In Homebrew_PCBs@yahoogroups.com, "Dave Mucha" <dave_mucha@y...> 
wrote:
> Read the home page of the list. <snip>

Here's a simpler way.

Use search and replace in a word processor to put a G0 in front of 
each X. (Search for "X" and replace with "G0 X")

Put the G81 canned drill cycle at the beginning:

G81 Z-.1 F30 R .1 (go down to -.1 at 30 units/minute, come back up) 
(to .1 above board at rapid rate.)

That's it. The g81 drilling cycle will repeat at each hole location 
in the list below.

Ballendo

P.S. You'd have your "safety block" (init commands) at the beginning 
of the file, and your "clean up and get out commands" at the end. But 
these would always be the same for a given board thickness and type. 
So with one search and replace, and three cut and pastes, you have 
converted an excellon file to a g code file. If you have different 
drill sizes, you'd add a search and replace to change the excellon 
toolchange to the Gcode toolchange.

P.P.S. You won't have to do even this with my machines; as they 
understand the excellon file directly.  

--- In Homebrew_PCBs@yahoogroups.com, "Dave Mucha" <dave_mucha@y...> 
wrote:

> here is a small snip of one of mine.
> 
> X009200Y004950
> X010400Y002250
> X009400Y002250
> X008400Y002250
> X007400Y002250
> X006400Y002250
> X005400Y002250
> X004400Y002250
> X003400Y002250
> X003400Y007050
> X004400Y007050
> X005400Y007050
> 
> 
> Cartesian coordinates in X and Y that exactly line up to the drill 
> holes.

--- In Homebrew_PCBs@yahoogroups.com, "Roy J. Tellason" 
<rtellason@b...> wrote:
<snip> 
>I haven't yet begun to explore the possibilities of various holddown 
>method. I can think of a few right off the top of my head,  those 
>being simple clamps,  some kind of a T-slot table (more clamps),  or 
>maybe a vacuum holddown setup.  What else is there?

Masking tape. By far the most common, and the most reliable, IMO. The 
registration is done with 1/8 or 3mm pins. You use the cnc to drill 
for them in the pcb AND the holes they fit into in the machine table 
(or a spoil board (recommended)). Just like a "real" board house 
tooling holes...

>Roy wrote of SMD's:
>That's someplace else I have no intention of going.  That stuff 
>does provide some benefit,  to be sure,  in terms of component 
>density and especially for manufacturing in large quantities,  but 
>again,  I don't see me getting into any of that sort of thing.

Common through hole parts are being discontinued with amazing 
regularity. Plan now your shift towards SMT and D or regret it 
later...

Hope this helps, 

Ballendo

On Wed, 28 Apr 2004 00:44:35 -0000, ballendo <ballendo@...> wrote:

> Stefan,
>
> I appreciate your comments, and the posts prior to this. I've snipped
> them to get us back on track...
>
> Okay, there's more info in those "short lines" than you think. For
> instance they have location obviously, but also a layer, and a
> filetype. And a width.


So you say:

I have to program either a tool that converts gerber to "dispenser file 
with 3d data"
or i have to edit the g-code myself and add the macros.
Second approach needs editing for each pcb you make.

There is no (free) solution where i can simply open my 2D gerber file and
then some already existing software adds the 3D information, letting me 
edit
parameters like depth and milling speed.


As i have no cnc experience i might just not get the point.
Currently it seems to me that the software would definitely take me longer 
to get
right than the hardware takes to be built.
I DO NOT LIKE that too much...


Is there any possibility i can test / try that without having a cnc mill?
When i have done cnc programming last i remember there was a simulator, 
which
showed you what it will be like.

Is there a free simulator? Which software do i need to download to try 
this?
Is there a good introductional page to cnc, showing the code commands and 
all?

thanks


ST

>In Homebrew_PCBs, "Dave Mucha" wrote:
> And THAT is why CNC-PCB_design was created !
> the non-technical, home brew drilling concept you can build from 
> stuff laying around to spending $100.00-200.00 in parts.

Hello,

This comment (above) caused me to do a quick check of something since 
I am a member of the mentioned group...

And sure enough, Dave Mucha has the blue star of list ownership!

A WHOLE lot of recent posts now make more sense, IMO...

Ballendo

P.S. My 500 buck machines use 3/4" steel ground and polished shafts; 
not drawer guides. Sometimes you get what you pay for... Drawer 
guides CAN work, I've built a few machines using them to learn their 
good and bad points. The machine shown at the group homepage will 
likely drill holes okay, but I'm suspect of it's milling capabilities 
and accuracy/repeatability. Based on my own experience with designs 
of this type. Starting in 1989...

>In Homebrew_PCBs, "Dave Mucha" wrote:
> And THAT is why CNC-PCB_design was created !
> 
> the non-technical, home brew drilling concept you can build from 
> stuff laying around to spending $100.00-200.00 in parts.
> 
> Dave

On Wed, 28 Apr 2004 08:54:54 -0000, ballendo <ballendo@...> wrote:

>> In Homebrew_PCBs, "Dave Mucha" wrote:
>> And THAT is why CNC-PCB_design was created !
>> the non-technical, home brew drilling concept you can build from
>> stuff laying around to spending $100.00-200.00 in parts.
>
> Hello,
>
> This comment (above) caused me to do a quick check of something since
> I am a member of the mentioned group...
>
> And sure enough, Dave Mucha has the blue star of list ownership!
>
> A WHOLE lot of recent posts now make more sense, IMO...
>
> Ballendo
>

Ya didn't know that ;-) Thought that was clear enough.....


I thought about this file editing for paste dispensing stuff...
I think i am perfectly capable of writing a tool in java or C++ that opens
a file, searches for each "hole" and inserting code there (some call to a 
macro).
Damn, what is this whole programming training good for if i never use it?
I think it would be easy to allow the user to enter the parameters of the 
"dispense
cycle".

What do you think, is it better to use a excellon file as a starting point 
or
a gerber file, what is more universal?
I can produce both, but future pcb software may not.
What is more common, and what is more easy to use?

I don't know how gerber stores the data, if i "draw" dots where i need the 
paste,
it it stored as "use aperture (round-1mm) at x y "?

Can you recommend a page where i can read up on gerber and excellon?

If using your canned drill cycle, where does one add the syringe stepper 
activation?


Well, i have to go now (you guess it - programming lessons ;-) i hate it
- unbelieveable how much programming you are forced to do if you study EE)


thanks

ST

P.S.: I am sorry that i hijacked that post, please reply with the proper 
Subject.

My apologies to the Homebrew PCB membership. Seems Dave has an axe to 
grind and has decided to do so publicly. Here's my public response...

Dave,

Thank you for sharing. My DIY-CNC list was started the last week of 
2003. There are now over 1000 members. About 300 people joined the 
first week. So someone likes the way I'm running it...

I have deleted some of your posts, because just like here, you seem 
to feel the need to moderate lists you don't own... NOT because of 
their topic.) Part of the reason the DIY-CNC list has grown so 
quickly is the fallout we receive from folks frustrated by the 
draconian moderation of CCED. But you know that, as you're the 
primary moderator there these days...)

Please refrain from speculation as to my intentions. You can simply 
ask, or read my very clearly stated aims, which are found in my posts.

Please stop moderating "other" peoples groups. 

Ballendo

P.S. If you prefer to talk to mechanical engineers about what 
electrical engineers/hobbyists want or need, that's your prerogative. 
In my many years of product design; I've learned to ask the 
CUSTOMERS. And then LISTEN to what they say. Ivory tower design 
rarely succeeds.


--- In Homebrew_PCBs@yahoogroups.com, "Dave Mucha" <dave_mucha@y...> 
wrote:
> I think Ballendo has a little desire to promote his unit and cheap 
> home-builts are part of his market, or competiton.  And it looks 
like the members on here are much more intereted in buying a unit 
whereas people on CNC-PCB_Design are there to build one.  Absolutly 
no customers there at all.

> 
> I know I have had posts on his list deleted for not being on topic 
> and they were really close to the topic, WAY closer than machine 
> design on this list.   If this was run under rules the way he runs 
> his lists, a lot of posts would have been deleted.  I think that is 
> why I got my dander up a little.

On Wednesday 28 April 2004 01:32 am, Dave Mucha wrote:

> Personally, if someone makes a dispenser, I'll be really peeved...
> if they DON'T post pictures !!!

Yeah.  In spite of however much somebody may write about what they've built, 
pictures sure *do* make it a lot easier to get a grip on things...

On Wednesday 28 April 2004 04:40 am, ballendo wrote:

> >Roy wrote of SMD's:
> >That's someplace else I have no intention of going.  That stuff
> >does provide some benefit,  to be sure,  in terms of component
> >density and especially for manufacturing in large quantities,  but
> >again,  I don't see me getting into any of that sort of thing.

> Common through hole parts are being discontinued with amazing
> regularity. Plan now your shift towards SMT and D or regret it
> later...

Not too likely.  I've scrapped enough stuff over the past several decades that 
I think it unlikely that I'll run out of parts any time soon.  As long as I'm 
willing to work with somewhat dated technology,  which I've decided I'm okay 
with. If it's that "new",  then I'll just get a board someplace and use it 
(as I'm doing with most of the computer stuff these days).

My eyes aren't what they used to be,  and as far as handling teeny little 
parts go,  my fingers seem to have gotten bigger or something.		:-)

So no,  for hobby stuff I don't plan on going there.

Dave Mucha wrote:
> --- In Homebrew_PCBs@yahoogroups.com, "ballendo" <ballendo@y...> 
> wrote:
> 
>>You're kidding right?
>>
>>We can talk for DAYS! about CuCl regeneration (which I could give a 
>>rat's @$$ about), but not about engineering a way to get folks more 
>>productive in the "other" aspects of pcb creation?
>
> This list does not cater to what you care about, it caters to the 
> topic as listed on the home page.
> 
> Read the home page of the list.  What are you reading into the 
> title ?  99.99% of all PCB's are made with chemicals.  This is the 
> RIGHT list for that.
> 
>>I think that the idea of "where" a certain topic might get "more 
>>play" is without merit. 
> 
> If you want machine design, you don't go to electical engineers.  If 
> you want to make a PCB you don't go to mechanical engineers and you 
> don't go to chemical engineers to find out how to drill holes.

CNC etching is just as OT as laser printer toner transfer and
printer/toner/inkjet/film discussions. Printers are CNC machines
too.

 > I
> imagine pcb design would get tossed off your list PDQ.  Why should 
> all other lists but this one have a specific topic as stated on the 
> home page?  Common sense should tell you where to go to get the best 
> response.  Of course it starts where the idea starts.  And if your 
> list started getting into the chemicals for making PCB's would you 
> want to discuss that or would you tell them to come here ?  ie: go to 
> where the experts are or people with similar interests.

The similar interests are to achieve the *end* result. I'd rather
see CNC discussions here than boring old laser printer crap which
i know is inconsistant between models and toner brands anyway. However,
i don't bitch about it OT.

>>FWIW, A "general" cnc machine list has NO IDEA what's required FOR 
>>pcb's. If you're gonna make a machine for pcb's, IMO you should 
>> talk with the folks who are making them!
> 
> I totally disagree.  How many on your list know the chemical 
> reactions of the stuff for making PCB's ?  Wrong list for that.  How 
> many on here know the stress and tortional calculations for 3/4 
> CRS ?  You don't ask an EE for mechanical designs and you don't as 
> ME's for Chemical properties and you don't ask a chemical engineer 
> how to drill holes.  (seems I already said that, but it is worth 
> repeating.)

PCB making crosses as many fields as you want. The list members come
from many fields. A mech.eng that wants to make PCBs might find
chemical etching OT just as much as i find laser printers OT.

>>since when does the "making" of the board stop 
>>before putting components on it? 
> 
> <snip>  
> 
>>Funny thing is, the idea is to allow someone LESS mechanically 
>>inclined to make (and complete) boards...
> 
> I agree, but you seem to advocating that less mechanically inclined 
> design your machines.  This list is not 90% mechanical and 10% 
> electronics.  just the opposite.

But i know *everything* electronic and have been learning more
things mechanical.

> I think the number of posts on the topics on here reflect the 
> background and experiance and direction of the list members.  90% 
> want to know about or handle chemcicals.

That's too restrictive. Discussions on lightbox design is just
as relevant.

> Look at any list, you go there to get answers or offer your help.  
> What part of this list description has anything to do with mechancial 
> machine design ?

Everything. PCBs are *mechanical*.

Regarding 90% of posts being about chemicals: yes, that was a thread
for quite a while as it covers quite a large territory and most here
are not chemists and so ask many questions. That doesn't mean people
only want to hear about it.

Regarding CNC/Pick 'n place/solder paste not being on topic: again I
point you to the home page:

Making, etching, soldering printed circuit boards yourself.

Any method allowed by law and local custom including photo, CAM, toner
transfer, direct printing, plotter, laundry pen, silkscreening,
toaster oven, paint stripper heat gun, etc.


Also look at Links. Note that I, the list-owner, made a folder for
mechanical etching/milling and placed links in it. This should be a clue.

I do agree that getting into the nitty gritty of modifying G-code may
be stretching the boundaries, as is 3D milling. -However- having
others play listcop is getting old. I do expect and welcome members
reminding others of the rules (I own or moderate a dozen lists!) but
finer points should be left to me. I can see this building up into
some real resentment between members.

Making printed circuit boards is a big topic. A huge topic. I said I
felt kind of funny when another list about CNC PCBs was started, but I
did not say anything because it -is- such a huge area to cover. I do
see some good reasons for another list to cover those more esoteric
and convoluted topics of strange codes and how to drive a stepper
motor or if servo motors are better. I'm not endorsing anyone's choice
of list.

So lets stop bickering and get back to talking about making PCBs.

Steve, list-owner and -only- moderator

Much simplier.

A note is that the G0  is a G and a ZERO, not the letter 'O'

My CNC software needs G00   (G-zero-zero) but this is still a 
simplier way.


A note of explanation is that the G00 is a go-to for the X/Y axis.

the drilling code
G81 Z-.1 F30 R .1 

is a program call to G81 (drill command)
Z-.1 is the Z axis command to go to -.1 
F30 is the feed rate to be 30% of the maximum speed of the axis.
R .1   is what is called a release point or a point above the part so 
the tool has clearance and moving the part will not let the drill hit 
any clamps or the part.

Another note is that the G81 will repeat ever time an new X/Y line is 
read.  so there is no need to jump to another for more drill 
instructions.

I currently use a bunch of different feeds and speeds on my holes so 
have created a table of different actions so the jump works best for 
me.

Dave







--- In Homebrew_PCBs@yahoogroups.com, "ballendo" <ballendo@y...> 
wrote:
> Here's a simpler way.
> 
> Use search and replace in a word processor to put a G0 in front of 
> each X. (Search for "X" and replace with "G0 X")
> 
> Put the G81 canned drill cycle at the beginning:
> 
> G81 Z-.1 F30 R .1 (go down to -.1 at 30 units/minute, come back up) 
> (to .1 above board at rapid rate.)
> 
> That's it. The g81 drilling cycle will repeat at each hole location 
> in the list below.
> 
> Ballendo
> 
> P.S. You'd have your "safety block" (init commands) at the 
beginning 
> of the file, and your "clean up and get out commands" at the end. 
But 
> these would always be the same for a given board thickness and 
type. 
> So with one search and replace, and three cut and pastes, you have 
> converted an excellon file to a g code file. If you have different 
> drill sizes, you'd add a search and replace to change the excellon 
> toolchange to the Gcode toolchange.
> 
> P.P.S. You won't have to do even this with my machines; as they 
> understand the excellon file directly.  
> 
> --- In Homebrew_PCBs@yahoogroups.com, "Dave Mucha" 
<dave_mucha@y...> 
> wrote:
> > here is a small snip of one of mine.
> > 
> > X009200Y004950
> > X010400Y002250
> > X009400Y002250
> > X008400Y002250
> > X007400Y002250
> > X006400Y002250
> > X005400Y002250
> > X004400Y002250
> > X003400Y002250
> > X003400Y007050
> > X004400Y007050
> > X005400Y007050
> > 
> > 
> > Cartesian coordinates in X and Y that exactly line up to the 
drill 
> > holes.

--- In Homebrew_PCBs@yahoogroups.com, "ballendo" <ballendo@y...> 
wrote:
> Dave,
> 
> Here's a copy of the FIRST line ON the home page:
> 
>   Making, etching, soldering printed circuit boards yourself.
> 
> See that third word there... 
> 
> Next, The list OWNER has approved this thread. And posted as such 
> already. Please respect what the LIST OWNER has said.


The whole question arose when a member posted that this is off topic, 
and I agreed.

Solder paste dispenser mechanical design is not even close.  
Mechanical etching is a legitimate and viable method of MAKING 
boards.   Yes the list owner will allow it, but that does not mean 
the responses will be the best for the person wanting to build the 
unit.

How about we agree that neither you nor I post on the solder paste or 
pic-and-place designs as we can see how the other members would 
answer this topic ?  (for a week or so ? )

I imagine it will get near zero play on this list.

Yes, you  are interested, and I am intereted, but this is not the 
mechanical design list for this topic.  Neither by the depth of 
member knoledge and experiance, nor the stated purpose of this list.

It IS a topic that would have dozens and dozens of professional 
responses on both your list and CCED.  Probably even have sites to 
people who have done it and personal experiances from those who 
already worked out the bugs.

I would make a screw plunger, that would need to be incremented as 
paste was diplaced from the syringe.  the A axis location would need 
to be advanced incrementally.  The next logical step is software and 
the next after that is troublshooting.  I think you mentioned a 
solenoid or compressed air ?  Neither of us have taken it to the 
final stage and worked out the bugs.

To keep it on here will eventally become a dis-service to the person 
making the unit.

And that, my friend, is why there are so many lists, and why both you 
and Steve have a couple dozen lists between you !

Dave

I think we are wearing this debate a little thin, so I'll let you 
have the last word.

self moderation is turned *ON*

<post sniped to make it easier to read, not all questions were 
addressed>

ballendo > > Okay, there's more info in those "short lines" than you 
think. For
> > instance they have location obviously, but also a layer, and a
> > filetype. And a width.
> 
> 
> So you say:
> 
> I have to program either a tool that converts gerber to "dispenser 
file > with 3d data"
> or i have to edit the g-code myself and add the macros.
> Second approach needs editing for each pcb you make.
> 
> There is no (free) solution where i can simply open my 2D gerber 
file and
> then some already existing software adds the 3D information, 
letting me 
> edit
> parameters like depth and milling speed.

This is correct.  I had mentioned that earlier on the thread.  It 
would not be a horrbily difficult matter for a programmer to read the 
NCDrill file and/or the Gerbers and 'remember' the last amount of 
paste in the tube and then output a G-code file to make that all 
happen.

I'm thinking we need to get a member who is fluent in programming to 
add that piece to the puzzle.

But, you are correct that you would need to have parameters like 
paste volume per pad would need to be a manual setting, preferrably 
one that is a manual potentiomer so you can do it on the fly.

depth can be a screw stop or a software setting, depending on the Z 
feed style.  Some are leadscrew, some are solenoid.   

And you are correct on speed.  there needs to be some hang time at 
the spot to allow the paste to dispense.



> 
> As i have no cnc experience i might just not get the point.
> Currently it seems to me that the software would definitely take me 
longer 
> to get
> right than the hardware takes to be built.
> I DO NOT LIKE that too much...


How right you are !

CNC is easy and you should at least play with one stepper to see it 
drive and how easy it can be.  tons of fun.

<snip> 
 
Dave

Dave Mucha wrote:
> --- In Homebrew_PCBs@yahoogroups.com, "ballendo" <ballendo@y...> 
> wrote:
> 
>>Dave,
>>
>>Here's a copy of the FIRST line ON the home page:
>>
>>  Making, etching, soldering printed circuit boards yourself.
>>
>>See that third word there... 
>>
>>Next, The list OWNER has approved this thread. And posted as such 
>>already. Please respect what the LIST OWNER has said.
> 
> 
> The whole question arose when a member posted that this is off topic, 
> and I agreed.
> 
> Solder paste dispenser mechanical design is not even close.  
> Mechanical etching is a legitimate and viable method of MAKING 
> boards.   Yes the list owner will allow it, but that does not mean 
> the responses will be the best for the person wanting to build the 
> unit.
> 
> How about we agree that neither you nor I post on the solder paste or 
> pic-and-place designs as we can see how the other members would 
> answer this topic ?  (for a week or so ? )
> 
> I imagine it will get near zero play on this list.

I happen to have designed the dispenser and pick-place
machine i want. I would have found such discussions interesting
here.

> Yes, you  are interested, and I am intereted, but this is not the 
> mechanical design list for this topic.  Neither by the depth of 
> member knoledge and experiance, nor the stated purpose of this list.

But what fun is a pcb without components.

> It IS a topic that would have dozens and dozens of professional 
> responses on both your list and CCED.  Probably even have sites to 
> people who have done it and personal experiances from those who 
> already worked out the bugs.
> 
> I would make a screw plunger, that would need to be incremented as 
> paste was diplaced from the syringe.  the A axis location would need 
> to be advanced incrementally.  The next logical step is software and 
> the next after that is troublshooting.  I think you mentioned a 
> solenoid or compressed air ?  Neither of us have taken it to the 
> final stage and worked out the bugs.

I have 2 solenoids, leavers, and plungers.

> To keep it on here will eventally become a dis-service to the person 
> making the unit.
> 
> And that, my friend, is why there are so many lists, and why both you 
> and Steve have a couple dozen lists between you !

When i wrote it is not completely correct that i know nothing about gerber
i meant that i have actually learned to program a cnc mill a few
years ago.
I do not remember much but i was able to guess what the G81 does.
I did write the code for the machine command by command then, not using
a drawing as a starting point.
This is why i understand the "cnc is controlled by file" part, and the
"layout is making a file" part but not exactly the "the files are not
the same and i must modify the layout file to use it on the cnc" part.



> Another note is that the G81 will repeat ever time an new X/Y line is
> read.  so there is no need to jump to another for more drill
> instructions.

I do not understand that part at all.
There is always a hole drilled when you read a x/y command?
You needn't call the cycle? seems strange...


>
> I currently use a bunch of different feeds and speeds on my holes so
> have created a table of different actions so the jump works best for
> me.
>

I think a few different cycles would be required for paste dispensing too.
might be good to have maybe 3 different "dot sizes".
Might even get impractical to do much more because it will not work good 
below and
above certain borders.

Do you think it makes any sense at all to try to place "beads" or is it
better to just make round "dots"? (and multiple if you need a line)

ST


  Dave

>
> This is correct.  I had mentioned that earlier on the thread.  It
> would not be a horrbily difficult matter for a programmer to read the
> NCDrill file and/or the Gerbers and 'remember' the last amount of
> paste in the tube and then output a G-code file to make that all
> happen.
>
> I'm thinking we need to get a member who is fluent in programming to
> add that piece to the puzzle.

I could do simple file editing, converting, i have done this before with 
java and c.
Would need to look up on graphic user interface, maybe someone wants
to do that. I would use a text interface for now.

>
> But, you are correct that you would need to have parameters like
> paste volume per pad would need to be a manual setting, preferrably
> one that is a manual potentiomer so you can do it on the fly.

A manual potentiometer would be one solution, i am not sure at all how
one could integrate the "4th axis" of the syringe motor in the G-code.
If one uses a pneumatic system the "potentiometer" would simply be the
pressure regulator.

>
> depth can be a screw stop or a software setting, depending on the Z
> feed style.  Some are leadscrew, some are solenoid.

I'd opt for a leadscrew type, as i want to enable drilling.
I think of a "gantry style" machine.


> And you are correct on speed.  there needs to be some hang time at
> the spot to allow the paste to dispense.

Yes, i do not think this is possible with a drill cycle at all, is it?
but i would assume one can make a own "paste cycle", like a macro
or what it is called (need to read up on that).

ST

Well, Russel,

maybe we can all stop bickering and rather use the valuable answer time
for generating useful posts.

You say you have made such a machine, tell me more.
Especially tell me what you do about your software, and
how the pneuamtic plunger works. i think i have all components
for a pneumatic system here (pump, pressure regulator, several solenoid 
valves).
I think it has some potential advantages and would like to give it a try.
I do not remember exactly how the air was coupled to the syringe, would it 
work
to simply apply the air pressure to the already existing plunger-back and 
seal
off the syringe housing?

thanks

ST

--- In Homebrew_PCBs@yahoogroups.com, Stefan Trethan 
<stefan_trethan@g...> wrote:
> On Wed, 28 Apr 2004 08:54:54 -0000, ballendo <ballendo@y...> wrote:
> 
> >> In Homebrew_PCBs, "Dave Mucha" wrote:
> >> And THAT is why CNC-PCB_design was created !
> >> the non-technical, home brew drilling concept you can build from
> >> stuff laying around to spending $100.00-200.00 in parts.
> >
> > Hello,
> >
> > This comment (above) caused me to do a quick check of something 
since
> > I am a member of the mentioned group...
> >
> > And sure enough, Dave Mucha has the blue star of list ownership!
> >
> > A WHOLE lot of recent posts now make more sense, IMO...
> >
> > Ballendo
> >


> 
> Ya didn't know that ;-) Thought that was clear enough.....



Gee, I thought that was common knoledge.

If you want the history, start with my post on 2788.

I asked how many people on this list want the machine stuff on here.

2 people.

I asked the moderator if he wanted the machine stuff on here.  

 no response.

I created  CNC-PCB_Design and didn't offer the list as an alternative 
but another resource for how to design and build.  but, that was 
almost 3 months after making the list and in response to a question 
on drill machine design.

and if you read the threads, others said they felt lost on CCED.

and...  if you read the posts, I agreed machine topics are better for 
machine lists.  the people on them have experiance and the technology.

Dave

thankfully, this whole thread is dying.  Knoledge somehow makes makes 
things easier to understand.

--- In Homebrew_PCBs@yahoogroups.com, Russell Shaw <rjshaw@n...> 

> > I would make a screw plunger, that would need to be incremented 
as 
> > paste was diplaced from the syringe.  the A axis location would 
need 
> > to be advanced incrementally.  The next logical step is software 
and 
> > the next after that is troublshooting.  I think you mentioned a 
> > solenoid or compressed air ?  Neither of us have taken it to the 
> > final stage and worked out the bugs.
> 
> I have 2 solenoids, leavers, and plungers.


Ok, Pleeeease  post stuff on your machine !

even if you don't want to give up the details in any large degree.

pictures ????
website ????

Dave

On Wednesday 28 April 2004 12:36 pm, Russell Shaw wrote:

> I happen to have designed the dispenser and pick-place machine i want.

Oh really?  Tell us more about it?  The latter,  in particular,  I'd be real 
interested in hearing a bit about just how that works,  and with what kind of 
parts.

> I would have found such discussions interesting here.

I would as well,  just to see what *can* be done if nothing else.  I look at a 
lot of what people are doing out there,  and while there's some of it that I 
may attempt at some point in time,  there's a fair amount of it I wouldn't 
ever consider.  I still find it interesting,  though.

<...>

> But what fun is a pcb without components.

Anyway!	:-)

Stefan Trethan wrote:
> Well, Russel,
> 
> maybe we can all stop bickering and rather use the valuable answer time
> for generating useful posts.
> 
> You say you have made such a machine, tell me more.
> Especially tell me what you do about your software, and
> how the pneuamtic plunger works. i think i have all components
> for a pneumatic system here (pump, pressure regulator, several solenoid 
> valves).
> I think it has some potential advantages and would like to give it a try.
> I do not remember exactly how the air was coupled to the syringe, would it 
> work
> to simply apply the air pressure to the already existing plunger-back and 
> seal
> off the syringe housing?
> 
> thanks

Hi,
I didn't intend to get anyones hopes high. I've designed it, and tested
most components/ideas in isolation, but haven't built the whole thing.
It has 12V/10A DC servos and is a design that is scalable from a 1m x 1m
area up to as large as needed such as 1.2m x 2.4m or larger (it's not
just for PCBs). Z axis travel is atleast 20cm. The software runs on a
linux pc, and sends basic shapes (path commands) to the microcontroller
on the machine. The paste dispenser uses metal plungers (i have a design
that also uses a single solenoid) instead of compressed air. The variable
viscosity due to batch variations, age, and temperature makes the flow of
solder paste unpredictable for a given air pressure. No plastic syringe will
be used. XYZ movement resolution will be 2.5um. Machine flexing will exceed
this, and the shaft/gearing backlash. Head movement at max servo speed will
be "quick" (60V servo drive ability), but i haven't got those numbers with
me atm. Motors are the most expensive part (AUS$120). Everything else is
*very* cheap. I wasn't going to say anything until i'd built it. I'm doing
the software between other jobs, so it'll take a while.

On Thu, 29 Apr 2004 18:36:48 +1000, Russell Shaw <rjshaw@...> 
wrote:

...

Ok, well, i see our design goals a a bit far apart.

However, one thing it there that i want to ask.
You say you don't use the plastic syringe, how do you fill any dispensing
mechanism "air bubble free"?
I did not consider buying paste in larger quantities or in jars mainly 
because
i think it is impossible to fill a syringe (or similar) without adding 
air, which
would in any case (maybe air pressure feed a bit less) influence the paste 
amount delivered,
and of course totally mess things up when the air bubble actually reaches 
the tip.



ST

Hello,

In a previous life I was a copier service tech. Prior to laser 
printers, but as most know now photocopiers use the same process.

We used to have a heck of a time with page registration, and I know 
the truth of Russell's comments below having adjusted many optics 
sets to<wards> 1:1...

So I'm wondering, Has anyone taken the time to see if he or she can 
get a good back to back registration on PAPER?

Because if you can't get the sheet of paper to reliably go through 
twice accurately; isn't it kinda unlikely that the mods to allow a 
board to go through are gonna improve things?

Next, IF it is possible and likely for the registration of two passes 
to line up with respect to the board edges: Has anyone held this up 
to a light and checked if all the holes line up all across the board?

For a eurocard sized board?

Thank you in advance,

Ballendo

P.S. I was really getting gung ho about direct etch resist printing--
even posted awhile back about the CD tray printer I saw at Fry's, and 
was talking to avecia about their resist and soldermask, when...

The above questions popped into my head. For me, if ya can't do two-
sided, it's a no-starter. SMD and SMT are not going away, and they 
require two-sided layouts more often than not (means always<G>)

I truly LOVE the idea. And with a print head on a cnc machine the 
registration problems go away... For others to take their place<G>

But if I'm gonna have to TT the second side anyways...


In Homebrew_PCBs@yahoogroups.com, Russell Shaw <rjshaw@n...> wrote:

> The similar interests are to achieve the *end* result. I'd rather
> see CNC discussions here than boring old laser printer crap which
> i know is inconsistant between models and toner brands anyway.

In Homebrew_PCBs, "Dave Mucha" <dave_mucha@y...> wrote:
<snip>
> F30 is the feed rate to be 30% of the maximum speed of the axis.
<snip>

Nope.

F30 means feed the axis or axes at a vector velocity of 30 UNITS PER 
MINUTE. Not a percentage of max speed. The units are either mm or 
inches. Some machines may also allow cm's to be used.

Any machine using percentage of max speed is not following the gerber 
or gcode "rules"...

Hope this helps,

Ballendo

On Thu, 29 Apr 2004 09:53:25 -0000, ballendo <ballendo@...> wrote:

> In Homebrew_PCBs, "Dave Mucha" <dave_mucha@y...> wrote:
> <snip>
>> F30 is the feed rate to be 30% of the maximum speed of the axis.
> <snip>
>
> Nope.
>
> F30 means feed the axis or axes at a vector velocity of 30 UNITS PER
> MINUTE. Not a percentage of max speed. The units are either mm or
> inches. Some machines may also allow cm's to be used.
>
> Any machine using percentage of max speed is not following the gerber
> or gcode "rules"...
>
> Hope this helps,
>
> Ballendo

Maybe dave has max. 100 Units per minute, then it would be the same.

Where can i read up on gerber?

ST

Hello,

>>Stefan wrote
>Dave Mucha wrote:
>>There is no (free) solution where i can simply open my 2D gerber 
>>file and then some already existing software adds the 3D 
>>information, letting me edit parameters like depth and milling 
>>speed.

>I'm thinking we need to get a member who is fluent in programming to 
>add that piece to the puzzle.

Got him already. Working on it.<G> For my machine users... Won't be 
free, but not too expensive either.

>'remember' the last amount of paste in the tube and then output a G-
>code file to make that all happen.

No need to "remember", just use incremental motion (G91) for the 
syringe leadscrew.

> But, you are correct that you would need to have parameters like 
> paste volume per pad would need to be a manual setting, preferrably 
> one that is a manual potentiomer so you can do it on the fly.

I don't agree. You need to have the ability to adjust the size 
of "dot". Once adjusted for a given machine and SMD footprint, it 
wouldn't need modifying "on the fly"... You're making it harder than 
it is...

> depth can be a screw stop or a software setting, depending on the Z 
> feed style.  Some are leadscrew, some are solenoid.   

Most commercial solder paste dispensers use the leadscrew model. The 
EFD unit has a remote leadscrew coupled to an axial plunger cable, 
like a lawn mower, or an aircraft throttle.

Ballendo

Russell,

Tell us about it!

Thank you in advance.

Ballendo

>In Homebrew_PCBs, Russell Shaw <rjshaw@n...> wrote:
> I happen to have designed the dispenser and pick-place
> machine i want. I would have found such discussions interesting
> here.

In Homebrew_PCBs, "Dave Mucha" <dave_mucha@y...> wrote:
>The next logical step is software and the next after that is 
>troublshooting.  I think you mentioned a solenoid or compressed 
>air ?  Neither of us have taken it to the final stage and worked out 
>the bugs.

That's news to me...

I'm still working on it, but I didn't START working on it when THIS 
thread commenced... Seems I'm a bit farther along than you think.

Dave, I've been making cnc machines for over 15 years. Some I can 
talk about; some I can't. Some are commercial; some are not. Some are 
for others; some are for me. This stuff isn't brand new to me...

Ballendo

In Homebrew_PCBs, Stefan Trethan <stefan_trethan@g...> wrote:
>>Another note is that the G81 will repeat ever time an new X/Y line 
>>is read.  so there is no need to jump to another for more drill
>>instructions.
> 
>I do not understand that part at all.
>There is always a hole drilled when you read a x/y command?
>You needn't call the cycle? seems strange...

Stefan,

The canned drill cycle is what's called "modal". That means that once 
it's called it "stays" in effect until something else replaces it. 
Like a subroutine. In fact, here's a bit of background on why they're 
called canned cycles:

Originally NC (no computer yet, so not C-NC) machines were programmed 
by paper tape with holes. Drilling a hole was a pretty standard thing 
to do, so after awhile folks started making tapes with nothing but 
the drill motions onit, repeated as many times as the tape length 
would allow. These were then kept in a CAN (to keep them away from 
the shop environment;paper, remember?) and spliced into the program 
whereever needed. When the computer came into the mix, and NC became 
CNC, this functionality was programmed, only now the splicing was 
done in computer memory (you know; HUGE 2K arrays of wires and 
ferrite beads<G>)

So anyway, after you've told the machine to make a certain series of 
motions to drill; it will repeat those motions at each new XY 
location you give it.

we're gonna use the down up motion of a drill cycle to depress a 
plunger and retract it a bit so it doesn't "drip". And I'm gonna put 
that drill cycle inside a macro so that the z axis goes down and up 
too...

>I think a few different cycles would be required for paste 
>dispensing too. 

Yes.

>Might be good to have maybe 3 different "dot sizes". Might even get 
>impractical to do much more because it will not work good below and 
>above certain borders.

Correct. But more than three may be needed, and certainly as many as 
are needed can be "made". 

Remember also that repeating a given size WITHOUT moving will make a 
bigger dot. So you could have a 1-dot,2-dot,4-dot, and get 8 sizes by 
combining them in binary fashion...

>Do you think it makes any sense at all to try to place "beads" or is 
>it better to just make round "dots"? (and multiple if you need a 
>line)

Dots are easier, and will probably do all that's needed. Beads are 
faster, but harder to do cheaply, IMO.

Hope this helps,

Ballendo

>In Homebrew_PCBs, Stefan Trethan <stefan_trethan@g...> wrote:
> 
>> And you are correct on speed.  there needs to be some hang time at
>> the spot to allow the paste to dispense.
> 
>Yes, i do not think this is possible with a drill cycle at all, is 
>it?

There are several different drill cycles. Drill with a dwell is one 
of them (G82). The dwell is set in milliseconds. For drilling this is 
used to be sure the drill makes complete revolution(s) at the bottom 
of a hole. We would use it for, as you guys are saying, "hang time".

>but i would assume one can make a own "paste cycle", like a macro
>or what it is called (need to read up on that).

Yes.

Ballendo

In fact, here's a bit of background on why they're
> called canned cycles:
>
> These were then kept in a CAN (to keep them away from
> the shop environment;paper, remember?)

I did wonder why canned cycles, but i certainly didn't expect THAT.

ST

> There are several different drill cycles. Drill with a dwell is one
> of them (G82). The dwell is set in milliseconds. For drilling this is
> used to be sure the drill makes complete revolution(s) at the bottom
> of a hole. We would use it for, as you guys are saying, "hang time".
>

ah, i see...

But is it really necessary to use drill cycles?
I mean if it prooves to be any awkward to use why not simply write a short
length of code along the lines of "move z down, activate syringe plunger, 
wait,
stop plunger, move up" you get the point... (see i still know no gerber 
because
you don't tell me where to read ;-) )

I don't think that would be much harder, must be possible to put it in a 
"macro"
and just call it every time you need?

You know, the canned cycles where invented to make things easier, if they 
make
things harder it might be better not to use the can but make a own piece
of code and "can it".

ST

ST,

read everything the manufacturers in a given field have in their 
literature. On their website... Trico and EFD are two of the "big 
players" in Dispensing.

Reading patents at www.uspto.gov can be VERY enlightening...

Search on "dispensing EFD"... (I think that's what I ended up using)

Note that many patents are expired and can be used freely. Many 
others can be licensed. Some will just give you ideas...

IMO, don't mess with copying an unexpired patent without permission! 
It is a FALLACY that you can USE a patented idea for your "own 
personal use" without selling it to anybody. While you might "get 
away" with this, it's NOT legal.  

Ballendo

P.S. many mfrs. will list their patent numbers on their website. Even 
expired ones. To "scare" you away... Read 'em.

--- In Homebrew_PCBs@yahoogroups.com, Stefan Trethan <> You say you 
have made such a machine, tell me more.
> Especially tell me what you do about your software, and
> how the pneuamtic plunger works. i think i have all components
> for a pneumatic system here (pump, pressure regulator, several 
solenoid 
> valves).
> I think it has some potential advantages and would like to give it 
a try.
> I do not remember exactly how the air was coupled to the syringe, 
would it 
> work
> to simply apply the air pressure to the already existing plunger-
back and 

> seal
> off the syringe housing?
> 
> thanks
> 
> ST

Ok i will do that..

But i want to read up on general cnc programming, g-code and stuff.
I could find some very specialized lists of command for certain machines 
but no
"G-code for dummies - a online introduction to CNC and gerber programming".

ST

On Thu, 29 Apr 2004 11:02:51 -0000, ballendo <ballendo@...> wrote:

> ST,
>
> read everything the manufacturers in a given field have in their
> literature. On their website... Trico and EFD are two of the "big
> players" in Dispensing.
>
> Reading patents at www.uspto.gov can be VERY enlightening...
>
> Search on "dispensing EFD"... (I think that's what I ended up using)
>
> Note that many patents are expired and can be used freely. Many
> others can be licensed. Some will just give you ideas...
>
> IMO, don't mess with copying an unexpired patent without permission!
> It is a FALLACY that you can USE a patented idea for your "own
> personal use" without selling it to anybody. While you might "get
> away" with this, it's NOT legal.
>
> Ballendo
>
> P.S. many mfrs. will list their patent numbers on their website. Even
> expired ones. To "scare" you away... Read 'em.

>In Homebrew_PCBs, Russell Shaw wrote:
>
> Head movement at max servo speed will be "quick" (60V servo drive 
> ability), but i haven't got those numbers with me atm. 

Russell, I'm interested in how you plan to feed the parts? And how 
you intend to orient them?

Sounds like it'll be a nice machine. I've excerpted the part above 
because it relates to the machine Mariss Freimanis uses for 
assembling his Geckodrives. That machine is a 14K parts per hour 
PNP...

I've been driving a fair bit lately, and was figuring out how fast it 
had to move to do that as I drove. Scary. Found out later it had 
three heads working simultaneously. STILL scary.<G>

Ballendo

P.S. lesseee, how many of my little hobby/small shop machines would 
it take for 14K parts per hour...<G> (Prob'ly still cheaper though!)

In Homebrew_PCBs, Stefan Trethan <stefan_trethan@g...> wrote:
> 
> But is it really necessary to use drill cycles?

Not necessary, but...

>I mean if it proves to be any awkward to use why not simply write a 
>short length of code along the lines of "move z down, activate 
>syringe plunger, wait, stop plunger, move up" you get the point... 

...you've just described the EXISTING drill cycle. Why not use it?

(see i still know no gerber because you don't tell me where to read ;-
) )

I have literally a hundred or so books on the subject of gcode and 
cnc. Along with many machine control programming and machine owners 
manuals. What I know I've learned over MANY years from MANY 
resources. I don't have a great resource for reading about gerbers 
online... But you could start at GC technologies, and see where it 
leads...

FWIW, gerber and Gcode are like a siter and brother. From the same 
family, very similar, but definitely different. And just like a 
sister and brother, who become MORE different as they grow; gerber 
and gcode have diverged further (farther? Troll?) as time has gone on.
You can read about gcode at several sites online. Best to just do a 
search of "G code" and "Gcode".

Peter Smid's book CNC Programming Handbook (about 35 bucks) is 
probably the best out right now, but just like all the others, it has 
mistakes you won't catch unless you're familiar with the subject. But 
it's worth its cost, IMO. Mike Lynch is another writer on gcode whose 
books are easy to follow. MANY cnc and Gcode books and online 
resources are TERRIBLE. 

> I don't think that would be much harder, must be possible to put it 
>in a "macro" and just call it every time you need?

Yes.

>You know, the canned cycles where invented to make things easier, if 
>they make things harder it might be better not to use the can but 
>make a own piece of code and "can it".

Yes. You have just described a macro<G>

Ballendo

Ask this question at the TurboCNC yahoo group... Download TurboCNC's 
manual at Dakeng.com

Download the manual for mach2 at www.artofcnc.ca

Ballendo

--- In Homebrew_PCBs@yahoogroups.com, Stefan Trethan 
<stefan_trethan@g...> wrote:
> Ok i will do that..
> 
> But i want to read up on general cnc programming, g-code and stuff.
> I could find some very specialized lists of command for certain 
machines 
> but no
> "G-code for dummies - a online introduction to CNC and gerber 
programming".
> 
> ST

Stefan Trethan wrote:
> On Thu, 29 Apr 2004 18:36:48 +1000, Russell Shaw <rjshaw@...> 
> wrote:
> 
> ...
> 
> Ok, well, i see our design goals a a bit far apart.

The reason for the large machine is that as well as for drilling
and milling, 100 reels of components uses up a bit of length.

> However, one thing it there that i want to ask.
> You say you don't use the plastic syringe, how do you fill any dispensing
> mechanism "air bubble free"?

I hadn't really considered it a problem, because the dispenser
software can predict the total volume dispensed, so that i can
refill it before air is sucked thru the mechanism. By squirting
new stuff in thru a tube where the tip is below the surface,
air is avoided. The other way is to connect a vacuum pump (or
simple reversed bicycle pump) and suck it down for 5-10s, which
pulls out small bubbles as very large bubbles under vacuum.
Letting the air back in causes any remaining bubbles to shrink
away. I have a simple tube fitting on it for that purpose if
needed. The valve arrangement avoids paste being sucked back
up the outlet.

> I did not consider buying paste in larger quantities or in jars mainly 
> because i think it is impossible to fill a syringe (or similar) without
 > adding air, which would in any case (maybe air pressure feed a bit less)
 > influence the paste amount delivered, and of course totally mess things
 > up when the air bubble actually reaches the tip.

My idea is to be able to push thru consistant volumes of paste, even
when the stuff is months old and would seize a normal syringe.

ballendo wrote:
>>In Homebrew_PCBs, Russell Shaw wrote:
>>
>>Head movement at max servo speed will be "quick" (60V servo drive 
>>ability), but i haven't got those numbers with me atm. 
> 
> Russell, I'm interested in how you plan to feed the parts? And how 
> you intend to orient them?

The tape from reels has a row of holes. The reels sits on a shaft
with a simple friction hub. The tape is fed between two rollers for
initial tape positioning. The tape base is pulled by a 3rd tooth
roller against another plain roller. Another pair of rollers does
the peeling of the top tape layer. The pickup is just a sawn-off
syringe needle with a flat end, and is spring-loaded vertically.
The vacuum (from a small diaphram pump) in the tube picks up the
component from the tape. After pickup, the component is held in
a simple device that can accurately orient the component in any
direction. The component is then placed.

> Sounds like it'll be a nice machine. I've excerpted the part above 
> because it relates to the machine Mariss Freimanis uses for 
> assembling his Geckodrives. That machine is a 14K parts per hour 
> PNP...
> 
> I've been driving a fair bit lately, and was figuring out how fast it 
> had to move to do that as I drove. Scary. Found out later it had 
> three heads working simultaneously. STILL scary.<G>

They use mnachine vision for orientation and compressed air to shoot
the components onto the paste.

> Ballendo
> 
> P.S. lesseee, how many of my little hobby/small shop machines would 
> it take for 14K parts per hour...<G> (Prob'ly still cheaper though!)

They need the high speed to justify the price they charge, and because
buyers are willing to pay for the speed. It's cheaper to design and
build your own machine these days (if you're a small outfit) if you
dedicate the machine to your own products that aren't likely to have a
huge volume. Machines are cheap at auctions, but support and accessories
still cost $$$.

The real problem with laser printers and backup is that the paper
physically changes size between printing the front and back side. This
is especially true in high-humidity areas (like here on the Wet Coast
:).

I've measured 1/8" shrinkage. If you run the paper through once (and say
print something off in a corner), you get MUCH less front-back
misregistration.

I was a Principal Software Developer at Creo (they manufacture
Computer-to-plate imagers for the printing industry), and one of the
products I was involved with was called Digital Blueline. This is
proofing software for doing backup proofs. We had the ability to shift
the back page to optimize alignment with the front page, and once the
paper was dehumidified, backup was pretty good.

So having said all that, I don't see any problem getting backup on a
media which doesn't change size, espcially if you have the ability to
apply an external shift. 

Doing backup with an inkjet should be MUCH easier.

--
Dave Hylands
Vancouver, BC, Canada
http://www.DaveHylands.com/

> -----Original Message-----
> From: ballendo [mailto:ballendo@...] 
> Sent: Thursday, April 29, 2004 2:31 AM
> To: Homebrew_PCBs@yahoogroups.com
> Subject: [Homebrew_PCBs] Re:laser jet consistency and registration
> 
> 
> Hello,
> 
> In a previous life I was a copier service tech. Prior to laser 
> printers, but as most know now photocopiers use the same process.
> 
> We used to have a heck of a time with page registration, and I know 
> the truth of Russell's comments below having adjusted many optics 
> sets to<wards> 1:1...
> 
> So I'm wondering, Has anyone taken the time to see if he or she can 
> get a good back to back registration on PAPER?
> 
> Because if you can't get the sheet of paper to reliably go through 
> twice accurately; isn't it kinda unlikely that the mods to allow a 
> board to go through are gonna improve things?
> 
> Next, IF it is possible and likely for the registration of two passes 
> to line up with respect to the board edges: Has anyone held this up 
> to a light and checked if all the holes line up all across the board?
> 
> For a eurocard sized board?
> 
> Thank you in advance,
> 
> Ballendo
> 
> P.S. I was really getting gung ho about direct etch resist 
> printing-- even posted awhile back about the CD tray printer 
> I saw at Fry's, and 
> was talking to avecia about their resist and soldermask, when...
> 
> The above questions popped into my head. For me, if ya can't 
> do two- sided, it's a no-starter. SMD and SMT are not going 
> away, and they 
> require two-sided layouts more often than not (means always<G>)
> 
> I truly LOVE the idea. And with a print head on a cnc machine the 
> registration problems go away... For others to take their place<G>
> 
> But if I'm gonna have to TT the second side anyways...
> 
> 
> In Homebrew_PCBs@yahoogroups.com, Russell Shaw <rjshaw@n...> wrote:
> > The similar interests are to achieve the *end* result. I'd 
> rather see 
> > CNC discussions here than boring old laser printer crap 
> which i know 
> > is inconsistant between models and toner brands anyway.
> 
> 
> 
> ------------------------ Yahoo! Groups Sponsor 
> ---------------------~--> Buy Ink Cartridges or Refill Kits 
> for your HP, Epson, Canon or Lexmark Printer at MyInks.com.  
> Free s/h on orders $50 or more to the US & Canada. 
http://www.c1tracking.com/l.asp?cid=5511
http://us.click.yahoo.com/mOAaAA/3exGAA/qnsNAA/bGYolB/TM
---------------------------------------------------------------------~->

Be sure to visit the group home and check for new Bookmarks and files:
http://groups.yahoo.com/group/Homebrew_PCBs 
Yahoo! Groups Links

Ballendo,
  I have done this on an HP 7150.  I have not tried with my epson color 800.
Someone mentioned using a scanner chassis to act as the X axis, instead  
of sheet
feeding.  You could put an aim point or jig on the bottom to hold the 
part e.g.: flush
left for top, flush right for bottom.  I know that most error I've seen 
has been from
page feed, not print head repeatability.  On the HP, I did get pretty 
good alignment,
but I have not looked at it lately.  I'll try and print a 2 sided board 
out tonight.
Richard

>From: "ballendo" <ballendo@...>
>Subject: Re:laser jet consistency and registration
><snip>
>So I'm wondering, Has anyone taken the time to see if he or she can 
>get a good back to back registration on PAPER?
>
>Because if you can't get the sheet of paper to reliably go through 
>twice accurately; isn't it kinda unlikely that the mods to allow a 
>board to go through are gonna improve things?
>
>Next, IF it is possible and likely for the registration of two passes 
>to line up with respect to the board edges: Has anyone held this up 
>to a light and checked if all the holes line up all across the board?
>  
>

We had a discussion about inkjet printing resist a while ago,
(actually we have it every few months with pretty much the same posts)
did anything come from it?

I'm still on the lookout for a higher-res laser printer for TT.
Don't want to buy one and tomorrow you come and tell me oh, hey,
i have done this PERFECT 1200DPI board by direct inkjet printing ;-).

Looking forward to hearing what is actually happening.

(And no, i don't think a 150dpi cd printer would do)

ST

>(And no, i don't think a 150dpi cd printer would do)

What about an epson 900 with cd feeder tray?

Is epsom ink resist? Could you put resist ink in a cartridge etc?


---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.639 / Virus Database: 408 - Release Date: 22/03/2004

Richard,

Thank you. I'll be interested in what you find... Especially the hole 
to hole registration on a 4x6 board...

Ballendo

--- In Homebrew_PCBs@yahoogroups.com, Richard Mustakos 
<rmustakos@a...> wrote:
> Ballendo,
>   I have done this on an HP 7150.  I have not tried with my epson 
color 800.
> Someone mentioned using a scanner chassis to act as the X axis, 
instead  
> of sheet
> feeding.  You could put an aim point or jig on the bottom to hold 
the 
> part e.g.: flush
> left for top, flush right for bottom.  I know that most error I've 
seen 
> has been from
> page feed, not print head repeatability.  On the HP, I did get 
pretty 
> good alignment,
> but I have not looked at it lately.  I'll try and print a 2 sided 
board 
> out tonight.
> Richard
> 
> >From: "ballendo" <ballendo@y...>
> >Subject: Re:laser jet consistency and registration
> ><snip>
> >So I'm wondering, Has anyone taken the time to see if he or she 
can 
> >get a good back to back registration on PAPER?
> >
> >Because if you can't get the sheet of paper to reliably go through 
> >twice accurately; isn't it kinda unlikely that the mods to allow a 
> >board to go through are gonna improve things?
> >
> >Next, IF it is possible and likely for the registration of two 
passes 
> >to line up with respect to the board edges: Has anyone held this 
up 
> >to a light and checked if all the holes line up all across the 
board?
> >  
> >

Has anybody tried the epson durabrite ink ? It is claimed to me water
and smudge resistant, so might be etch resistant too. I was looking
for printers that support these ink cartridges, there are ones that
are not too expensive, but I can't get one with a straight through
path here. The ones that have straight through path doesn't support
durabrite cartridges :-(

-- Rajesh

--- In Homebrew_PCBs@yahoogroups.com, "Lez" <lez.briddon@n...> wrote:

>  
> 
> 
> >(And no, i don't think a 150dpi cd printer would do)
> 
> What about an epson 900 with cd feeder tray?
> 
> Is epsom ink resist? Could you put resist ink in a cartridge etc?
> 
> 
> ---
> Outgoing mail is certified Virus Free.
> Checked by AVG anti-virus system (http://www.grisoft.com).
> Version: 6.0.639 / Virus Database: 408 - Release Date: 22/03/2004

On Fri, 30 Apr 2004 09:32:08 -0000, Rajesh LG <rajeshlg@...> wrote:

> Has anybody tried the epson durabrite ink ? It is claimed to me water
> and smudge resistant, so might be etch resistant too. I was looking
> for printers that support these ink cartridges, there are ones that
> are not too expensive, but I can't get one with a straight through
> path here. The ones that have straight through path doesn't support
> durabrite cartridges :-(
>
> -- Rajesh
>

If i have a prooven concept that it works i would simply modify the printer
to have a straight through path.

Look out for "cd/dvd printer conversion" you can find guides how to do it 
and
you also can find out which printers have a straight path.
Some have a "envelope" slot at the back, but normally drag the paper in 
 from below.
these might just work.


Please if you do research collect all relevant pages in a folder in the 
links section.

ST

> Look out for "cd/dvd printer conversion" you can find guides how to 
do it 
> and
> you also can find out which printers have a straight path.


CD printer ?  I am guessing they have ink that would adhere to the 
plastic surface much like for copper ?

Most inkjets only need to get the ink near paper and the paper sucks 
it in like a sponge.

This may be a good cross over ink.

Dave

these are special "print on me" cds, coated similar to ohp transparency.
I think maybe the wax printers can use uncoated...

ST

On Fri, 30 Apr 2004 15:13:41 -0000, Dave Mucha <dave_mucha@...> 
wrote:

>
>> Look out for "cd/dvd printer conversion" you can find guides how to
> do it
>> and
>> you also can find out which printers have a straight path.
>
>
> CD printer ?  I am guessing they have ink that would adhere to the
> plastic surface much like for copper ?
>
> Most inkjets only need to get the ink near paper and the paper sucks
> it in like a sponge.
>
> This may be a good cross over ink.
>
> Dave
>
>
>
>
>
> Be sure to visit the group home and check for new Bookmarks and files:
> http://groups.yahoo.com/group/Homebrew_PCBs
> Yahoo! Groups Links
>
>
>
>
>

On the few double sided TT I have done, I put registration marks well 
off the board edges and line up the two pages on those first.  then 
slip the board between.

The registration is so close that the annular rings on both sides are 
fully formed.   The drill holes are often not as precise.  the 
registraton is withing a few mils.   

HP-LJ-4 is what I have used.

One of the unknowns about copiers and printers is that they are not 
really 1:1 in both X and Y.  When you flip your board and re-print, 
the board would be in the same top of sheet orientatation.  if you 
flipped AND rotated 90 degrees, your alignement would be worse.


Dave




--- In Homebrew_PCBs@yahoogroups.com, "Dave Hylands" <dhylands@b...> 
wrote:
> The real problem with laser printers and backup is that the paper
> physically changes size between printing the front and back side. 
This
> is especially true in high-humidity areas (like here on the Wet 
Coast
> :).
> 
> I've measured 1/8" shrinkage. If you run the paper through once 
(and say
> print something off in a corner), you get MUCH less front-back
> misregistration.
> 
> I was a Principal Software Developer at Creo (they manufacture
> Computer-to-plate imagers for the printing industry), and one of the
> products I was involved with was called Digital Blueline. This is
> proofing software for doing backup proofs. We had the ability to 
shift
> the back page to optimize alignment with the front page, and once 
the
> paper was dehumidified, backup was pretty good.
> 
> So having said all that, I don't see any problem getting backup on a
> media which doesn't change size, espcially if you have the ability 
to
> apply an external shift. 
> 
> Doing backup with an inkjet should be MUCH easier.
> 
> --
> Dave Hylands
> Vancouver, BC, Canada
> http://www.DaveHylands.com/
> 
> > -----Original Message-----
> > From: ballendo [mailto:ballendo@y...] 
> > Sent: Thursday, April 29, 2004 2:31 AM
> > To: Homebrew_PCBs@yahoogroups.com
> > Subject: [Homebrew_PCBs] Re:laser jet consistency and registration
> > 
> > 
> > Hello,
> > 
> > In a previous life I was a copier service tech. Prior to laser 
> > printers, but as most know now photocopiers use the same process.
> > 
> > We used to have a heck of a time with page registration, and I 
know 
> > the truth of Russell's comments below having adjusted many optics 
> > sets to<wards> 1:1...
> > 
> > So I'm wondering, Has anyone taken the time to see if he or she 
can 
> > get a good back to back registration on PAPER?
> > 
> > Because if you can't get the sheet of paper to reliably go 
through 
> > twice accurately; isn't it kinda unlikely that the mods to allow 
a 
> > board to go through are gonna improve things?
> > 
> > Next, IF it is possible and likely for the registration of two 
passes 
> > to line up with respect to the board edges: Has anyone held this 
up 
> > to a light and checked if all the holes line up all across the 
board?
> > 
> > For a eurocard sized board?
> > 
> > Thank you in advance,
> > 
> > Ballendo
> > 
> > P.S. I was really getting gung ho about direct etch resist 
> > printing-- even posted awhile back about the CD tray printer 
> > I saw at Fry's, and 
> > was talking to avecia about their resist and soldermask, when...
> > 
> > The above questions popped into my head. For me, if ya can't 
> > do two- sided, it's a no-starter. SMD and SMT are not going 
> > away, and they 
> > require two-sided layouts more often than not (means always<G>)
> > 
> > I truly LOVE the idea. And with a print head on a cnc machine the 
> > registration problems go away... For others to take their place<G>
> > 
> > But if I'm gonna have to TT the second side anyways...
> > 
> > 
> > In Homebrew_PCBs@yahoogroups.com, Russell Shaw <rjshaw@n...> 
wrote:
> > > The similar interests are to achieve the *end* result. I'd 
> > rather see 
> > > CNC discussions here than boring old laser printer crap 
> > which i know 
> > > is inconsistant between models and toner brands anyway.
> > 
> > 
> > 
> > ------------------------ Yahoo! Groups Sponsor 
> > ---------------------~--> Buy Ink Cartridges or Refill Kits 
> > for your HP, Epson, Canon or Lexmark Printer at MyInks.com.  
> > Free s/h on orders $50 or more to the US & Canada. 
> http://www.c1tracking.com/l.asp?cid=5511
> http://us.click.yahoo.com/mOAaAA/3exGAA/qnsNAA/bGYolB/TM
> --------------------------------------------------------------------
-~->
> 
> Be sure to visit the group home and check for new Bookmarks and 
files:
> http://groups.yahoo.com/group/Homebrew_PCBs 
> Yahoo! Groups Links

--- In Homebrew_PCBs@yahoogroups.com, Stefan Trethan 
<stefan_trethan@g...> wrote:
> these are special "print on me" cds, coated similar to ohp 
transparency.
> I think maybe the wax printers can use uncoated...
> 
> ST


a quick google shows some CD printers are thermal wax, others are for 
special CD's.   I am just starting to look so don't know all the in's 
and outs as yet.

Dave

BTW Stefan, what part of the world do you call home ?  I know you are 
not in the US, but I was just curious.

Dave



> 
> On Fri, 30 Apr 2004 15:13:41 -0000, Dave Mucha <dave_mucha@y...> 
> wrote:
> 
> >
> >> Look out for "cd/dvd printer conversion" you can find guides how 
to
> > do it
> >> and
> >> you also can find out which printers have a straight path.
> >
> >
> > CD printer ?  I am guessing they have ink that would adhere to the
> > plastic surface much like for copper ?
> >
> > Most inkjets only need to get the ink near paper and the paper 
sucks
> > it in like a sponge.
> >
> > This may be a good cross over ink.
> >
> > Dave
> >
> >
> >
> >
> >
> > Be sure to visit the group home and check for new Bookmarks and 
files:

> > http://groups.yahoo.com/group/Homebrew_PCBs
> > Yahoo! Groups Links
> >
> >
> >
> >
> >

> BTW Stefan, what part of the world do you call home ?  I know you are
> not in the US, but I was just curious.
>
> Dave
>
>


.at is Austria.

ST

FYI, Solder dispenser, 1st bid @ 9.99 (not me selling)
http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&category=58293&item=3813015150&rd=1&ssPageName=WDVW
Looks like it would not be a hard hack if it's still working.
Richard